d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
9ad221a558
freebsd-dev/sys/dev/iwm/if_iwm.c

6279 lines
170 KiB
C
Raw Normal View History

iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/* $OpenBSD: if_iwm.c,v 1.42 2015/05/30 02:49:23 deraadt Exp $ */
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* Copyright (c) 2014 genua mbh <info@genua.de>
* Copyright (c) 2014 Fixup Software Ltd.
*
* Permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
/*-
* Based on BSD-licensed source modules in the Linux iwlwifi driver,
* which were used as the reference documentation for this implementation.
*
* Driver version we are currently based off of is
* Linux 3.14.3 (tag id a2df521e42b1d9a23f620ac79dbfe8655a8391dd)
*
***********************************************************************
*
* This file is provided under a dual BSD/GPLv2 license. When using or
* redistributing this file, you may do so under either license.
*
* GPL LICENSE SUMMARY
*
* Copyright(c) 2007 - 2013 Intel Corporation. All rights reserved.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of version 2 of the GNU General Public License as
* published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110,
* USA
*
* The full GNU General Public License is included in this distribution
* in the file called COPYING.
*
* Contact Information:
* Intel Linux Wireless <ilw@linux.intel.com>
* Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
*
*
* BSD LICENSE
*
* Copyright(c) 2005 - 2013 Intel Corporation. All rights reserved.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
* * Neither the name Intel Corporation nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
/*-
* Copyright (c) 2007-2010 Damien Bergamini <damien.bergamini@free.fr>
*
* Permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
iwm: add 'opt_wlan.h' include into source files. Include net80211 options file, so IEEE80211_DEBUG_REFCNT option will be handled correctly.
2016-05-19 23:31:00 +00:00
#include "opt_wlan.h"
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
#include <sys/param.h>
#include <sys/bus.h>
#include <sys/conf.h>
#include <sys/endian.h>
#include <sys/firmware.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/mutex.h>
#include <sys/module.h>
#include <sys/proc.h>
#include <sys/rman.h>
#include <sys/socket.h>
#include <sys/sockio.h>
#include <sys/sysctl.h>
#include <sys/linker.h>
#include <machine/bus.h>
#include <machine/endian.h>
#include <machine/resource.h>
#include <dev/pci/pcivar.h>
#include <dev/pci/pcireg.h>
#include <net/bpf.h>
#include <net/if.h>
#include <net/if_var.h>
#include <net/if_arp.h>
#include <net/if_dl.h>
#include <net/if_media.h>
#include <net/if_types.h>
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/if_ether.h>
#include <netinet/ip.h>
#include <net80211/ieee80211_var.h>
#include <net80211/ieee80211_regdomain.h>
#include <net80211/ieee80211_ratectl.h>
#include <net80211/ieee80211_radiotap.h>
iwm: use the proper include path for iwm headers.
2015-08-08 21:08:35 +00:00
#include <dev/iwm/if_iwmreg.h>
#include <dev/iwm/if_iwmvar.h>
#include <dev/iwm/if_iwm_debug.h>
#include <dev/iwm/if_iwm_util.h>
#include <dev/iwm/if_iwm_binding.h>
#include <dev/iwm/if_iwm_phy_db.h>
#include <dev/iwm/if_iwm_mac_ctxt.h>
#include <dev/iwm/if_iwm_phy_ctxt.h>
#include <dev/iwm/if_iwm_time_event.h>
#include <dev/iwm/if_iwm_power.h>
#include <dev/iwm/if_iwm_scan.h>
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm: use the proper include path for iwm headers.
2015-08-08 21:08:35 +00:00
#include <dev/iwm/if_iwm_pcie_trans.h>
[iwm] add LED blinking for iwm hardware that supports it. Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD, Linux iwlwifi/mvm
2016-06-02 04:42:28 +00:00
#include <dev/iwm/if_iwm_led.h>
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
const uint8_t iwm_nvm_channels[] = {
/* 2.4 GHz */
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
/* 5 GHz */
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
36, 40, 44, 48, 52, 56, 60, 64,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
100, 104, 108, 112, 116, 120, 124, 128, 132, 136, 140, 144,
149, 153, 157, 161, 165
};
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
_Static_assert(nitems(iwm_nvm_channels) <= IWM_NUM_CHANNELS,
"IWM_NUM_CHANNELS is too small");
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
const uint8_t iwm_nvm_channels_8000[] = {
/* 2.4 GHz */
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
/* 5 GHz */
36, 40, 44, 48, 52, 56, 60, 64, 68, 72, 76, 80, 84, 88, 92,
96, 100, 104, 108, 112, 116, 120, 124, 128, 132, 136, 140, 144,
149, 153, 157, 161, 165, 169, 173, 177, 181
};
_Static_assert(nitems(iwm_nvm_channels_8000) <= IWM_NUM_CHANNELS_8000,
"IWM_NUM_CHANNELS_8000 is too small");
#define IWM_NUM_2GHZ_CHANNELS 14
#define IWM_N_HW_ADDR_MASK 0xF
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* XXX For now, there's simply a fixed set of rate table entries
* that are populated.
*/
const struct iwm_rate {
uint8_t rate;
uint8_t plcp;
} iwm_rates[] = {
{ 2, IWM_RATE_1M_PLCP },
{ 4, IWM_RATE_2M_PLCP },
{ 11, IWM_RATE_5M_PLCP },
{ 22, IWM_RATE_11M_PLCP },
{ 12, IWM_RATE_6M_PLCP },
{ 18, IWM_RATE_9M_PLCP },
{ 24, IWM_RATE_12M_PLCP },
{ 36, IWM_RATE_18M_PLCP },
{ 48, IWM_RATE_24M_PLCP },
{ 72, IWM_RATE_36M_PLCP },
{ 96, IWM_RATE_48M_PLCP },
{ 108, IWM_RATE_54M_PLCP },
};
#define IWM_RIDX_CCK 0
#define IWM_RIDX_OFDM 4
#define IWM_RIDX_MAX (nitems(iwm_rates)-1)
#define IWM_RIDX_IS_CCK(_i_) ((_i_) < IWM_RIDX_OFDM)
#define IWM_RIDX_IS_OFDM(_i_) ((_i_) >= IWM_RIDX_OFDM)
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
struct iwm_nvm_section {
uint16_t length;
uint8_t *data;
};
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int iwm_store_cscheme(struct iwm_softc *, const uint8_t *, size_t);
static int iwm_firmware_store_section(struct iwm_softc *,
enum iwm_ucode_type,
const uint8_t *, size_t);
static int iwm_set_default_calib(struct iwm_softc *, const void *);
static void iwm_fw_info_free(struct iwm_fw_info *);
static int iwm_read_firmware(struct iwm_softc *, enum iwm_ucode_type);
static void iwm_dma_map_addr(void *, bus_dma_segment_t *, int, int);
static int iwm_dma_contig_alloc(bus_dma_tag_t, struct iwm_dma_info *,
bus_size_t, bus_size_t);
static void iwm_dma_contig_free(struct iwm_dma_info *);
static int iwm_alloc_fwmem(struct iwm_softc *);
static int iwm_alloc_sched(struct iwm_softc *);
static int iwm_alloc_kw(struct iwm_softc *);
static int iwm_alloc_ict(struct iwm_softc *);
static int iwm_alloc_rx_ring(struct iwm_softc *, struct iwm_rx_ring *);
[iwm] Add bit-polling in Rx-DMA init code path. Taken-From: OpenBSD (if_iwm.c r1.80) Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD ed35558754288911048cb607e57c688273ebd8d4
2016-06-02 05:00:52 +00:00
static void iwm_disable_rx_dma(struct iwm_softc *);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static void iwm_reset_rx_ring(struct iwm_softc *, struct iwm_rx_ring *);
static void iwm_free_rx_ring(struct iwm_softc *, struct iwm_rx_ring *);
static int iwm_alloc_tx_ring(struct iwm_softc *, struct iwm_tx_ring *,
int);
static void iwm_reset_tx_ring(struct iwm_softc *, struct iwm_tx_ring *);
static void iwm_free_tx_ring(struct iwm_softc *, struct iwm_tx_ring *);
static void iwm_enable_interrupts(struct iwm_softc *);
static void iwm_restore_interrupts(struct iwm_softc *);
static void iwm_disable_interrupts(struct iwm_softc *);
static void iwm_ict_reset(struct iwm_softc *);
static int iwm_allow_mcast(struct ieee80211vap *, struct iwm_softc *);
static void iwm_stop_device(struct iwm_softc *);
static void iwm_mvm_nic_config(struct iwm_softc *);
static int iwm_nic_rx_init(struct iwm_softc *);
static int iwm_nic_tx_init(struct iwm_softc *);
static int iwm_nic_init(struct iwm_softc *);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
static int iwm_enable_txq(struct iwm_softc *, int, int, int);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int iwm_post_alive(struct iwm_softc *);
static int iwm_nvm_read_chunk(struct iwm_softc *, uint16_t, uint16_t,
uint16_t, uint8_t *, uint16_t *);
static int iwm_nvm_read_section(struct iwm_softc *, uint16_t, uint8_t *,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
uint16_t *, size_t);
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
static uint32_t iwm_eeprom_channel_flags(uint16_t);
static void iwm_add_channel_band(struct iwm_softc *,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
struct ieee80211_channel[], int, int *, int, size_t,
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
const uint8_t[]);
static void iwm_init_channel_map(struct ieee80211com *, int, int *,
struct ieee80211_channel[]);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int iwm_parse_nvm_data(struct iwm_softc *, const uint16_t *,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
const uint16_t *, const uint16_t *,
const uint16_t *, const uint16_t *,
const uint16_t *);
static void iwm_set_hw_address_8000(struct iwm_softc *,
struct iwm_nvm_data *,
const uint16_t *, const uint16_t *);
static int iwm_get_sku(const struct iwm_softc *, const uint16_t *,
const uint16_t *);
static int iwm_get_nvm_version(const struct iwm_softc *, const uint16_t *);
static int iwm_get_radio_cfg(const struct iwm_softc *, const uint16_t *,
const uint16_t *);
static int iwm_get_n_hw_addrs(const struct iwm_softc *,
const uint16_t *);
static void iwm_set_radio_cfg(const struct iwm_softc *,
struct iwm_nvm_data *, uint32_t);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int iwm_parse_nvm_sections(struct iwm_softc *,
struct iwm_nvm_section *);
static int iwm_nvm_init(struct iwm_softc *);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
static int iwm_firmware_load_sect(struct iwm_softc *, uint32_t,
const uint8_t *, uint32_t);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int iwm_firmware_load_chunk(struct iwm_softc *, uint32_t,
const uint8_t *, uint32_t);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
static int iwm_load_firmware_7000(struct iwm_softc *, enum iwm_ucode_type);
static int iwm_load_cpu_sections_8000(struct iwm_softc *,
struct iwm_fw_sects *, int , int *);
static int iwm_load_firmware_8000(struct iwm_softc *, enum iwm_ucode_type);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int iwm_load_firmware(struct iwm_softc *, enum iwm_ucode_type);
static int iwm_start_fw(struct iwm_softc *, enum iwm_ucode_type);
static int iwm_send_tx_ant_cfg(struct iwm_softc *, uint8_t);
static int iwm_send_phy_cfg_cmd(struct iwm_softc *);
static int iwm_mvm_load_ucode_wait_alive(struct iwm_softc *,
enum iwm_ucode_type);
static int iwm_run_init_mvm_ucode(struct iwm_softc *, int);
static int iwm_rx_addbuf(struct iwm_softc *, int, int);
static int iwm_mvm_calc_rssi(struct iwm_softc *, struct iwm_rx_phy_info *);
static int iwm_mvm_get_signal_strength(struct iwm_softc *,
struct iwm_rx_phy_info *);
static void iwm_mvm_rx_rx_phy_cmd(struct iwm_softc *,
struct iwm_rx_packet *,
struct iwm_rx_data *);
[iwm] fix up RSSI calculations for both scan results and normal RX operations. * hard code a noise floor of -96 for now. The noise floor update code returns some "interesting" values that I can't map to anything useful right now. * Ensure a default noise floor is set - otherwise the initial scan results have a noise floor of '0'. * Fix up the RSSI calculation to be correctly relative to the noise floor. The RSSI routines return an absolute value in dBm - so fix this up. * Cap RSSI values appropriately. * Ensure we pass in a 1/2 dB unit value in to net80211. Tested: * Intel 7260, STA mode iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: hw rev 0x140, fw ver 16.242414.0, address xx:xx:xx:xx:xx:xx
2016-09-18 05:07:18 +00:00
static int iwm_get_noise(struct iwm_softc *sc,
const struct iwm_mvm_statistics_rx_non_phy *);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static void iwm_mvm_rx_rx_mpdu(struct iwm_softc *, struct iwm_rx_packet *,
struct iwm_rx_data *);
iwm: convert to ieee80211_tx_complete() Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D4755
2016-01-03 10:10:11 +00:00
static int iwm_mvm_rx_tx_cmd_single(struct iwm_softc *,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
struct iwm_rx_packet *,
struct iwm_node *);
static void iwm_mvm_rx_tx_cmd(struct iwm_softc *, struct iwm_rx_packet *,
struct iwm_rx_data *);
static void iwm_cmd_done(struct iwm_softc *, struct iwm_rx_packet *);
#if 0
static void iwm_update_sched(struct iwm_softc *, int, int, uint8_t,
uint16_t);
#endif
static const struct iwm_rate *
iwm_tx_fill_cmd(struct iwm_softc *, struct iwm_node *,
[iwm] use rate control info from the node txrates; use mgmtrate for EAPOL frames This changes the transmit rate control code to do a few things: * use fixed rates (mcast, ucast, mgmt) where required. * Don't use a hard-coded 11a or 11bg rate for non-data frames - use what net80211 says we should use. * use mgmtrate for EAPOL frames. Reviewed by: avos Differential Revision: https://reviews.freebsd.org/D7994
2016-09-21 20:56:10 +00:00
struct mbuf *, struct iwm_tx_cmd *);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int iwm_tx(struct iwm_softc *, struct mbuf *,
struct ieee80211_node *, int);
static int iwm_raw_xmit(struct ieee80211_node *, struct mbuf *,
const struct ieee80211_bpf_params *);
[iwm] Improve reliability of iwm_release on disassociation a bit. * We need to first call ivp->iv_newstate(), to enqueue the deauth/deassoc mgmt frame, then flush the tx queue, before actually calling iwm_release(). * cycling a wlan connection via wpa_cli frontend to wpa_supplicant, by issuing disconnect and reconnect commandos works quite well. (There is still an issue when disconnecting/reconnecting too quickly) * Reassociating or roaming via wpa_supplicant is still broken. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7943
2016-09-19 22:36:30 +00:00
static int iwm_mvm_flush_tx_path(struct iwm_softc *sc,
uint32_t tfd_msk, uint32_t flags);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int iwm_mvm_send_add_sta_cmd_status(struct iwm_softc *,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
struct iwm_mvm_add_sta_cmd_v7 *,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
int *);
static int iwm_mvm_sta_send_to_fw(struct iwm_softc *, struct iwm_node *,
int);
static int iwm_mvm_add_sta(struct iwm_softc *, struct iwm_node *);
static int iwm_mvm_update_sta(struct iwm_softc *, struct iwm_node *);
static int iwm_mvm_add_int_sta_common(struct iwm_softc *,
struct iwm_int_sta *,
const uint8_t *, uint16_t, uint16_t);
static int iwm_mvm_add_aux_sta(struct iwm_softc *);
static int iwm_mvm_update_quotas(struct iwm_softc *, struct iwm_node *);
static int iwm_auth(struct ieee80211vap *, struct iwm_softc *);
static int iwm_assoc(struct ieee80211vap *, struct iwm_softc *);
static int iwm_release(struct iwm_softc *, struct iwm_node *);
static struct ieee80211_node *
iwm_node_alloc(struct ieee80211vap *,
const uint8_t[IEEE80211_ADDR_LEN]);
static void iwm_setrates(struct iwm_softc *, struct iwm_node *);
static int iwm_media_change(struct ifnet *);
static int iwm_newstate(struct ieee80211vap *, enum ieee80211_state, int);
static void iwm_endscan_cb(void *, int);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
static void iwm_mvm_fill_sf_command(struct iwm_softc *,
struct iwm_sf_cfg_cmd *,
struct ieee80211_node *);
static int iwm_mvm_sf_config(struct iwm_softc *, enum iwm_sf_state);
static int iwm_send_bt_init_conf(struct iwm_softc *);
static int iwm_send_update_mcc_cmd(struct iwm_softc *, const char *);
static void iwm_mvm_tt_tx_backoff(struct iwm_softc *, uint32_t);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int iwm_init_hw(struct iwm_softc *);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
static void iwm_init(struct iwm_softc *);
static void iwm_start(struct iwm_softc *);
static void iwm_stop(struct iwm_softc *);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static void iwm_watchdog(void *);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
static void iwm_parent(struct ieee80211com *);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
#ifdef IWM_DEBUG
static const char *
iwm_desc_lookup(uint32_t);
static void iwm_nic_error(struct iwm_softc *);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
static void iwm_nic_umac_error(struct iwm_softc *);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
#endif
static void iwm_notif_intr(struct iwm_softc *);
static void iwm_intr(void *);
static int iwm_attach(device_t);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
static int iwm_is_valid_ether_addr(uint8_t *);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static void iwm_preinit(void *);
static int iwm_detach_local(struct iwm_softc *sc, int);
static void iwm_init_task(void *);
static void iwm_radiotap_attach(struct iwm_softc *);
static struct ieee80211vap *
iwm_vap_create(struct ieee80211com *,
const char [IFNAMSIZ], int,
enum ieee80211_opmode, int,
const uint8_t [IEEE80211_ADDR_LEN],
const uint8_t [IEEE80211_ADDR_LEN]);
static void iwm_vap_delete(struct ieee80211vap *);
static void iwm_scan_start(struct ieee80211com *);
static void iwm_scan_end(struct ieee80211com *);
static void iwm_update_mcast(struct ieee80211com *);
static void iwm_set_channel(struct ieee80211com *);
static void iwm_scan_curchan(struct ieee80211_scan_state *, unsigned long);
static void iwm_scan_mindwell(struct ieee80211_scan_state *);
static int iwm_detach(device_t);
/*
* Firmware parser.
*/
static int
iwm_store_cscheme(struct iwm_softc *sc, const uint8_t *data, size_t dlen)
{
const struct iwm_fw_cscheme_list *l = (const void *)data;
if (dlen < sizeof(*l) ||
dlen < sizeof(l->size) + l->size * sizeof(*l->cs))
return EINVAL;
/* we don't actually store anything for now, always use s/w crypto */
return 0;
}
static int
iwm_firmware_store_section(struct iwm_softc *sc,
enum iwm_ucode_type type, const uint8_t *data, size_t dlen)
{
struct iwm_fw_sects *fws;
struct iwm_fw_onesect *fwone;
if (type >= IWM_UCODE_TYPE_MAX)
return EINVAL;
if (dlen < sizeof(uint32_t))
return EINVAL;
fws = &sc->sc_fw.fw_sects[type];
if (fws->fw_count >= IWM_UCODE_SECT_MAX)
return EINVAL;
fwone = &fws->fw_sect[fws->fw_count];
/* first 32bit are device load offset */
memcpy(&fwone->fws_devoff, data, sizeof(uint32_t));
/* rest is data */
fwone->fws_data = data + sizeof(uint32_t);
fwone->fws_len = dlen - sizeof(uint32_t);
fws->fw_count++;
return 0;
}
[iwm] Use IWM_DEFAULT_SCAN_CHANNELS define as default for sc_capa_n_scan_channels. Approved by: adrian (mentor) Obtained from: Linux iwlwifi Differential Revision: https://reviews.freebsd.org/D7938
2016-09-19 22:23:46 +00:00
#define IWM_DEFAULT_SCAN_CHANNELS 40
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* iwlwifi: iwl-drv.c */
struct iwm_tlv_calib_data {
uint32_t ucode_type;
struct iwm_tlv_calib_ctrl calib;
} __packed;
static int
iwm_set_default_calib(struct iwm_softc *sc, const void *data)
{
const struct iwm_tlv_calib_data *def_calib = data;
uint32_t ucode_type = le32toh(def_calib->ucode_type);
if (ucode_type >= IWM_UCODE_TYPE_MAX) {
device_printf(sc->sc_dev,
"Wrong ucode_type %u for default "
"calibration.\n", ucode_type);
return EINVAL;
}
sc->sc_default_calib[ucode_type].flow_trigger =
def_calib->calib.flow_trigger;
sc->sc_default_calib[ucode_type].event_trigger =
def_calib->calib.event_trigger;
return 0;
}
static void
iwm_fw_info_free(struct iwm_fw_info *fw)
{
iwm: store pointer for 'struct firmware' instead of 'size_t' and 'void *' pair. Approved by: adrian (mentor) Obtained from: DragonFlyBSD Differential Revision: https://reviews.freebsd.org/D4765
2016-01-04 21:03:01 +00:00
firmware_put(fw->fw_fp, FIRMWARE_UNLOAD);
fw->fw_fp = NULL;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* don't touch fw->fw_status */
memset(fw->fw_sects, 0, sizeof(fw->fw_sects));
}
static int
iwm_read_firmware(struct iwm_softc *sc, enum iwm_ucode_type ucode_type)
{
struct iwm_fw_info *fw = &sc->sc_fw;
const struct iwm_tlv_ucode_header *uhdr;
struct iwm_ucode_tlv tlv;
enum iwm_ucode_tlv_type tlv_type;
const struct firmware *fwp;
const uint8_t *data;
int error = 0;
size_t len;
if (fw->fw_status == IWM_FW_STATUS_DONE &&
ucode_type != IWM_UCODE_TYPE_INIT)
return 0;
while (fw->fw_status == IWM_FW_STATUS_INPROGRESS)
msleep(&sc->sc_fw, &sc->sc_mtx, 0, "iwmfwp", 0);
fw->fw_status = IWM_FW_STATUS_INPROGRESS;
iwm: store pointer for 'struct firmware' instead of 'size_t' and 'void *' pair. Approved by: adrian (mentor) Obtained from: DragonFlyBSD Differential Revision: https://reviews.freebsd.org/D4765
2016-01-04 21:03:01 +00:00
if (fw->fw_fp != NULL)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_fw_info_free(fw);
/*
* Load firmware into driver memory.
iwm: store pointer for 'struct firmware' instead of 'size_t' and 'void *' pair. Approved by: adrian (mentor) Obtained from: DragonFlyBSD Differential Revision: https://reviews.freebsd.org/D4765
2016-01-04 21:03:01 +00:00
* fw_fp will be set.
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
*/
IWM_UNLOCK(sc);
fwp = firmware_get(sc->sc_fwname);
iwm: store pointer for 'struct firmware' instead of 'size_t' and 'void *' pair. Approved by: adrian (mentor) Obtained from: DragonFlyBSD Differential Revision: https://reviews.freebsd.org/D4765
2016-01-04 21:03:01 +00:00
IWM_LOCK(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (fwp == NULL) {
device_printf(sc->sc_dev,
"could not read firmware %s (error %d)\n",
sc->sc_fwname, error);
goto out;
}
iwm: store pointer for 'struct firmware' instead of 'size_t' and 'void *' pair. Approved by: adrian (mentor) Obtained from: DragonFlyBSD Differential Revision: https://reviews.freebsd.org/D4765
2016-01-04 21:03:01 +00:00
fw->fw_fp = fwp;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/* (Re-)Initialize default values. */
sc->sc_capaflags = 0;
[iwm] Use IWM_DEFAULT_SCAN_CHANNELS define as default for sc_capa_n_scan_channels. Approved by: adrian (mentor) Obtained from: Linux iwlwifi Differential Revision: https://reviews.freebsd.org/D7938
2016-09-19 22:23:46 +00:00
sc->sc_capa_n_scan_channels = IWM_DEFAULT_SCAN_CHANNELS;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
memset(sc->sc_enabled_capa, 0, sizeof(sc->sc_enabled_capa));
memset(sc->sc_fw_mcc, 0, sizeof(sc->sc_fw_mcc));
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* Parse firmware contents
*/
iwm: store pointer for 'struct firmware' instead of 'size_t' and 'void *' pair. Approved by: adrian (mentor) Obtained from: DragonFlyBSD Differential Revision: https://reviews.freebsd.org/D4765
2016-01-04 21:03:01 +00:00
uhdr = (const void *)fw->fw_fp->data;
if (*(const uint32_t *)fw->fw_fp->data != 0
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
|| le32toh(uhdr->magic) != IWM_TLV_UCODE_MAGIC) {
device_printf(sc->sc_dev, "invalid firmware %s\n",
sc->sc_fwname);
error = EINVAL;
goto out;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
snprintf(sc->sc_fwver, sizeof(sc->sc_fwver), "%d.%d (API ver %d)",
IWM_UCODE_MAJOR(le32toh(uhdr->ver)),
IWM_UCODE_MINOR(le32toh(uhdr->ver)),
IWM_UCODE_API(le32toh(uhdr->ver)));
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
data = uhdr->data;
iwm: store pointer for 'struct firmware' instead of 'size_t' and 'void *' pair. Approved by: adrian (mentor) Obtained from: DragonFlyBSD Differential Revision: https://reviews.freebsd.org/D4765
2016-01-04 21:03:01 +00:00
len = fw->fw_fp->datasize - sizeof(*uhdr);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
while (len >= sizeof(tlv)) {
size_t tlv_len;
const void *tlv_data;
memcpy(&tlv, data, sizeof(tlv));
tlv_len = le32toh(tlv.length);
tlv_type = le32toh(tlv.type);
len -= sizeof(tlv);
data += sizeof(tlv);
tlv_data = data;
if (len < tlv_len) {
device_printf(sc->sc_dev,
"firmware too short: %zu bytes\n",
len);
error = EINVAL;
goto parse_out;
}
switch ((int)tlv_type) {
case IWM_UCODE_TLV_PROBE_MAX_LEN:
if (tlv_len < sizeof(uint32_t)) {
device_printf(sc->sc_dev,
"%s: PROBE_MAX_LEN (%d) < sizeof(uint32_t)\n",
__func__,
(int) tlv_len);
error = EINVAL;
goto parse_out;
}
sc->sc_capa_max_probe_len
= le32toh(*(const uint32_t *)tlv_data);
/* limit it to something sensible */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (sc->sc_capa_max_probe_len >
IWM_SCAN_OFFLOAD_PROBE_REQ_SIZE) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_FIRMWARE_TLV,
"%s: IWM_UCODE_TLV_PROBE_MAX_LEN "
"ridiculous\n", __func__);
error = EINVAL;
goto parse_out;
}
break;
case IWM_UCODE_TLV_PAN:
if (tlv_len) {
device_printf(sc->sc_dev,
"%s: IWM_UCODE_TLV_PAN: tlv_len (%d) > 0\n",
__func__,
(int) tlv_len);
error = EINVAL;
goto parse_out;
}
sc->sc_capaflags |= IWM_UCODE_TLV_FLAGS_PAN;
break;
case IWM_UCODE_TLV_FLAGS:
if (tlv_len < sizeof(uint32_t)) {
device_printf(sc->sc_dev,
"%s: IWM_UCODE_TLV_FLAGS: tlv_len (%d) < sizeof(uint32_t)\n",
__func__,
(int) tlv_len);
error = EINVAL;
goto parse_out;
}
/*
* Apparently there can be many flags, but Linux driver
* parses only the first one, and so do we.
*
* XXX: why does this override IWM_UCODE_TLV_PAN?
* Intentional or a bug? Observations from
* current firmware file:
* 1) TLV_PAN is parsed first
* 2) TLV_FLAGS contains TLV_FLAGS_PAN
* ==> this resets TLV_PAN to itself... hnnnk
*/
sc->sc_capaflags = le32toh(*(const uint32_t *)tlv_data);
break;
case IWM_UCODE_TLV_CSCHEME:
if ((error = iwm_store_cscheme(sc,
tlv_data, tlv_len)) != 0) {
device_printf(sc->sc_dev,
"%s: iwm_store_cscheme(): returned %d\n",
__func__,
error);
goto parse_out;
}
break;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
case IWM_UCODE_TLV_NUM_OF_CPU: {
uint32_t num_cpu;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (tlv_len != sizeof(uint32_t)) {
device_printf(sc->sc_dev,
"%s: IWM_UCODE_TLV_NUM_OF_CPU: tlv_len (%d) < sizeof(uint32_t)\n",
__func__,
(int) tlv_len);
error = EINVAL;
goto parse_out;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
num_cpu = le32toh(*(const uint32_t *)tlv_data);
if (num_cpu < 1 || num_cpu > 2) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
"%s: Driver supports only 1 or 2 CPUs\n",
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
__func__);
error = EINVAL;
goto parse_out;
}
break;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
case IWM_UCODE_TLV_SEC_RT:
if ((error = iwm_firmware_store_section(sc,
IWM_UCODE_TYPE_REGULAR, tlv_data, tlv_len)) != 0) {
device_printf(sc->sc_dev,
"%s: IWM_UCODE_TYPE_REGULAR: iwm_firmware_store_section() failed; %d\n",
__func__,
error);
goto parse_out;
}
break;
case IWM_UCODE_TLV_SEC_INIT:
if ((error = iwm_firmware_store_section(sc,
IWM_UCODE_TYPE_INIT, tlv_data, tlv_len)) != 0) {
device_printf(sc->sc_dev,
"%s: IWM_UCODE_TYPE_INIT: iwm_firmware_store_section() failed; %d\n",
__func__,
error);
goto parse_out;
}
break;
case IWM_UCODE_TLV_SEC_WOWLAN:
if ((error = iwm_firmware_store_section(sc,
IWM_UCODE_TYPE_WOW, tlv_data, tlv_len)) != 0) {
device_printf(sc->sc_dev,
"%s: IWM_UCODE_TYPE_WOW: iwm_firmware_store_section() failed; %d\n",
__func__,
error);
goto parse_out;
}
break;
case IWM_UCODE_TLV_DEF_CALIB:
if (tlv_len != sizeof(struct iwm_tlv_calib_data)) {
device_printf(sc->sc_dev,
"%s: IWM_UCODE_TLV_DEV_CALIB: tlv_len (%d) < sizeof(iwm_tlv_calib_data) (%d)\n",
__func__,
(int) tlv_len,
(int) sizeof(struct iwm_tlv_calib_data));
error = EINVAL;
goto parse_out;
}
if ((error = iwm_set_default_calib(sc, tlv_data)) != 0) {
device_printf(sc->sc_dev,
"%s: iwm_set_default_calib() failed: %d\n",
__func__,
error);
goto parse_out;
}
break;
case IWM_UCODE_TLV_PHY_SKU:
if (tlv_len != sizeof(uint32_t)) {
error = EINVAL;
device_printf(sc->sc_dev,
"%s: IWM_UCODE_TLV_PHY_SKU: tlv_len (%d) < sizeof(uint32_t)\n",
__func__,
(int) tlv_len);
goto parse_out;
}
sc->sc_fw_phy_config =
le32toh(*(const uint32_t *)tlv_data);
break;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
case IWM_UCODE_TLV_API_CHANGES_SET: {
const struct iwm_ucode_api *api;
if (tlv_len != sizeof(*api)) {
error = EINVAL;
goto parse_out;
}
api = (const struct iwm_ucode_api *)tlv_data;
/* Flags may exceed 32 bits in future firmware. */
if (le32toh(api->api_index) > 0) {
device_printf(sc->sc_dev,
"unsupported API index %d\n",
le32toh(api->api_index));
goto parse_out;
}
sc->sc_ucode_api = le32toh(api->api_flags);
break;
}
case IWM_UCODE_TLV_ENABLED_CAPABILITIES: {
const struct iwm_ucode_capa *capa;
int idx, i;
if (tlv_len != sizeof(*capa)) {
error = EINVAL;
goto parse_out;
}
capa = (const struct iwm_ucode_capa *)tlv_data;
idx = le32toh(capa->api_index);
[iwm] Fix off-by-one check in iwm_read_firmware(). This fixes a potential buffer overrun in the firmware parsing code. Reported by: Coverity Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7931
2016-09-19 22:11:34 +00:00
if (idx >= howmany(IWM_NUM_UCODE_TLV_CAPA, 32)) {
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
device_printf(sc->sc_dev,
"unsupported API index %d\n", idx);
goto parse_out;
}
for (i = 0; i < 32; i++) {
if ((le32toh(capa->api_capa) & (1U << i)) == 0)
continue;
setbit(sc->sc_enabled_capa, i + (32 * idx));
}
break;
}
case 48: /* undocumented TLV */
case IWM_UCODE_TLV_SDIO_ADMA_ADDR:
case IWM_UCODE_TLV_FW_GSCAN_CAPA:
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* ignore, not used by current driver */
break;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
case IWM_UCODE_TLV_SEC_RT_USNIFFER:
if ((error = iwm_firmware_store_section(sc,
IWM_UCODE_TYPE_REGULAR_USNIFFER, tlv_data,
tlv_len)) != 0)
goto parse_out;
break;
case IWM_UCODE_TLV_N_SCAN_CHANNELS:
if (tlv_len != sizeof(uint32_t)) {
error = EINVAL;
goto parse_out;
}
sc->sc_capa_n_scan_channels =
le32toh(*(const uint32_t *)tlv_data);
break;
case IWM_UCODE_TLV_FW_VERSION:
if (tlv_len != sizeof(uint32_t) * 3) {
error = EINVAL;
goto parse_out;
}
snprintf(sc->sc_fwver, sizeof(sc->sc_fwver),
"%d.%d.%d",
le32toh(((const uint32_t *)tlv_data)[0]),
le32toh(((const uint32_t *)tlv_data)[1]),
le32toh(((const uint32_t *)tlv_data)[2]));
break;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
default:
device_printf(sc->sc_dev,
"%s: unknown firmware section %d, abort\n",
__func__, tlv_type);
error = EINVAL;
goto parse_out;
}
len -= roundup(tlv_len, 4);
data += roundup(tlv_len, 4);
}
KASSERT(error == 0, ("unhandled error"));
parse_out:
if (error) {
device_printf(sc->sc_dev, "firmware parse error %d, "
"section type %d\n", error, tlv_type);
}
if (!(sc->sc_capaflags & IWM_UCODE_TLV_FLAGS_PM_CMD_SUPPORT)) {
device_printf(sc->sc_dev,
"device uses unsupported power ops\n");
error = ENOTSUP;
}
out:
if (error) {
fw->fw_status = IWM_FW_STATUS_NONE;
iwm: store pointer for 'struct firmware' instead of 'size_t' and 'void *' pair. Approved by: adrian (mentor) Obtained from: DragonFlyBSD Differential Revision: https://reviews.freebsd.org/D4765
2016-01-04 21:03:01 +00:00
if (fw->fw_fp != NULL)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_fw_info_free(fw);
} else
fw->fw_status = IWM_FW_STATUS_DONE;
wakeup(&sc->sc_fw);
return error;
}
/*
* DMA resource routines
*/
static void
iwm_dma_map_addr(void *arg, bus_dma_segment_t *segs, int nsegs, int error)
{
if (error != 0)
return;
KASSERT(nsegs == 1, ("too many DMA segments, %d should be 1", nsegs));
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
*(bus_addr_t *)arg = segs[0].ds_addr;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
static int
iwm_dma_contig_alloc(bus_dma_tag_t tag, struct iwm_dma_info *dma,
bus_size_t size, bus_size_t alignment)
{
int error;
dma->tag = NULL;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
dma->map = NULL;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
dma->size = size;
[iwm] Fix iwm_dma_contig_free(). dma->map is always NULL here. * When bus_dmamem_alloc is used, the bus_dmamap_t is usually set to NULL, so we were never actually freeing any dma memory allocations done via iwm_dma_contig_alloc(). So we should check dma->vaddr instead of dma->map here. * Also, the dmamap is actually supposed to be invalidated as part of bus_dmamem_free(), so bus_dmamap_destroy() is never needed here. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Obtained from: DragonflyBSD git ef2b29a7ba6ca8a9d2c82ab591c0622227ff84cb
2016-06-23 01:00:13 +00:00
dma->vaddr = NULL;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
error = bus_dma_tag_create(tag, alignment,
0, BUS_SPACE_MAXADDR_32BIT, BUS_SPACE_MAXADDR, NULL, NULL, size,
Remove BUS_DMA_NOWAIT from bus_dma_tag_create() invocations as it's no valid flag there.
2015-10-21 06:23:57 +00:00
1, size, 0, NULL, NULL, &dma->tag);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (error != 0)
goto fail;
error = bus_dmamem_alloc(dma->tag, (void **)&dma->vaddr,
BUS_DMA_NOWAIT | BUS_DMA_ZERO | BUS_DMA_COHERENT, &dma->map);
if (error != 0)
goto fail;
error = bus_dmamap_load(dma->tag, dma->map, dma->vaddr, size,
iwm_dma_map_addr, &dma->paddr, BUS_DMA_NOWAIT);
[iwm] Fix iwm_dma_contig_free(). dma->map is always NULL here. * When bus_dmamem_alloc is used, the bus_dmamap_t is usually set to NULL, so we were never actually freeing any dma memory allocations done via iwm_dma_contig_alloc(). So we should check dma->vaddr instead of dma->map here. * Also, the dmamap is actually supposed to be invalidated as part of bus_dmamem_free(), so bus_dmamap_destroy() is never needed here. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Obtained from: DragonflyBSD git ef2b29a7ba6ca8a9d2c82ab591c0622227ff84cb
2016-06-23 01:00:13 +00:00
if (error != 0) {
bus_dmamem_free(dma->tag, dma->vaddr, dma->map);
dma->vaddr = NULL;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
goto fail;
[iwm] Fix iwm_dma_contig_free(). dma->map is always NULL here. * When bus_dmamem_alloc is used, the bus_dmamap_t is usually set to NULL, so we were never actually freeing any dma memory allocations done via iwm_dma_contig_alloc(). So we should check dma->vaddr instead of dma->map here. * Also, the dmamap is actually supposed to be invalidated as part of bus_dmamem_free(), so bus_dmamap_destroy() is never needed here. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Obtained from: DragonflyBSD git ef2b29a7ba6ca8a9d2c82ab591c0622227ff84cb
2016-06-23 01:00:13 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
bus_dmamap_sync(dma->tag, dma->map, BUS_DMASYNC_PREWRITE);
return 0;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
fail:
iwm_dma_contig_free(dma);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
}
static void
iwm_dma_contig_free(struct iwm_dma_info *dma)
{
[iwm] Fix iwm_dma_contig_free(). dma->map is always NULL here. * When bus_dmamem_alloc is used, the bus_dmamap_t is usually set to NULL, so we were never actually freeing any dma memory allocations done via iwm_dma_contig_alloc(). So we should check dma->vaddr instead of dma->map here. * Also, the dmamap is actually supposed to be invalidated as part of bus_dmamem_free(), so bus_dmamap_destroy() is never needed here. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Obtained from: DragonflyBSD git ef2b29a7ba6ca8a9d2c82ab591c0622227ff84cb
2016-06-23 01:00:13 +00:00
if (dma->vaddr != NULL) {
bus_dmamap_sync(dma->tag, dma->map,
BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
bus_dmamap_unload(dma->tag, dma->map);
bus_dmamem_free(dma->tag, dma->vaddr, dma->map);
dma->vaddr = NULL;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
if (dma->tag != NULL) {
bus_dma_tag_destroy(dma->tag);
dma->tag = NULL;
}
}
/* fwmem is used to load firmware onto the card */
static int
iwm_alloc_fwmem(struct iwm_softc *sc)
{
/* Must be aligned on a 16-byte boundary. */
return iwm_dma_contig_alloc(sc->sc_dmat, &sc->fw_dma,
sc->sc_fwdmasegsz, 16);
}
/* tx scheduler rings. not used? */
static int
iwm_alloc_sched(struct iwm_softc *sc)
{
/* TX scheduler rings must be aligned on a 1KB boundary. */
[iwm] When stopping TX DMA, wait for all channels at once. * Makes the TX DMA stopping more similar to Linux code, and potentially a bit faster. Also, output an error message when TX DMA idling fails. Taken-From: Linux iwlwifi Tested: * AC3165, STA mode Approved by: adrian (mentor) Obtained from: DragonFlyBSD git 2ee486ddff973ac552ff787c17e8d83e8ae0f24c Differential Revision: https://reviews.freebsd.org/D7325
2016-07-27 20:51:31 +00:00
return iwm_dma_contig_alloc(sc->sc_dmat, &sc->sched_dma,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
nitems(sc->txq) * sizeof(struct iwm_agn_scd_bc_tbl), 1024);
}
/* keep-warm page is used internally by the card. see iwl-fh.h for more info */
static int
iwm_alloc_kw(struct iwm_softc *sc)
{
return iwm_dma_contig_alloc(sc->sc_dmat, &sc->kw_dma, 4096, 4096);
}
/* interrupt cause table */
static int
iwm_alloc_ict(struct iwm_softc *sc)
{
return iwm_dma_contig_alloc(sc->sc_dmat, &sc->ict_dma,
IWM_ICT_SIZE, 1<<IWM_ICT_PADDR_SHIFT);
}
static int
iwm_alloc_rx_ring(struct iwm_softc *sc, struct iwm_rx_ring *ring)
{
bus_size_t size;
int i, error;
ring->cur = 0;
/* Allocate RX descriptors (256-byte aligned). */
size = IWM_RX_RING_COUNT * sizeof(uint32_t);
error = iwm_dma_contig_alloc(sc->sc_dmat, &ring->desc_dma, size, 256);
if (error != 0) {
device_printf(sc->sc_dev,
"could not allocate RX ring DMA memory\n");
goto fail;
}
ring->desc = ring->desc_dma.vaddr;
/* Allocate RX status area (16-byte aligned). */
error = iwm_dma_contig_alloc(sc->sc_dmat, &ring->stat_dma,
sizeof(*ring->stat), 16);
if (error != 0) {
device_printf(sc->sc_dev,
"could not allocate RX status DMA memory\n");
goto fail;
}
ring->stat = ring->stat_dma.vaddr;
/* Create RX buffer DMA tag. */
error = bus_dma_tag_create(sc->sc_dmat, 1, 0,
BUS_SPACE_MAXADDR_32BIT, BUS_SPACE_MAXADDR, NULL, NULL,
Remove BUS_DMA_NOWAIT from bus_dma_tag_create() invocations as it's no valid flag there.
2015-10-21 06:23:57 +00:00
IWM_RBUF_SIZE, 1, IWM_RBUF_SIZE, 0, NULL, NULL, &ring->data_dmat);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (error != 0) {
device_printf(sc->sc_dev,
"%s: could not create RX buf DMA tag, error %d\n",
__func__, error);
goto fail;
}
[iwm] Fix up busdma use in the RX path When allocating a new mbuf or bus_dmamap_load()-ing it fails, we can just keep the old mbuf since we are dropping that packet anyway. Instead of doing bus_dmamap_create() and bus_dmamap_destroy() all the time, create an extra bus_dmamap_t which we can use to safely try bus_dmamap_load()-ing the new mbuf. On success we just swap the spare bus_dmamap_t with the data->map of that ring entry. Tested: Tested with Intel AC7260, verified with vmstat -m that new kernel no longer visibly leaks memory from the M_DEVBUF malloc type. Before, leakage was 1KB every few seconds while ping(8)-ing over the wlan connection. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re@ Obtained from: DragonflyBSD.git cc440b26818b5dfdd9af504d71c1b0e6522b53ef Differential Revision: https://reviews.freebsd.org/D6742
2016-06-13 00:13:20 +00:00
/* Allocate spare bus_dmamap_t for iwm_rx_addbuf() */
error = bus_dmamap_create(ring->data_dmat, 0, &ring->spare_map);
if (error != 0) {
device_printf(sc->sc_dev,
"%s: could not create RX buf DMA map, error %d\n",
__func__, error);
goto fail;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* Allocate and map RX buffers.
*/
for (i = 0; i < IWM_RX_RING_COUNT; i++) {
[iwm] Fix up busdma use in the RX path When allocating a new mbuf or bus_dmamap_load()-ing it fails, we can just keep the old mbuf since we are dropping that packet anyway. Instead of doing bus_dmamap_create() and bus_dmamap_destroy() all the time, create an extra bus_dmamap_t which we can use to safely try bus_dmamap_load()-ing the new mbuf. On success we just swap the spare bus_dmamap_t with the data->map of that ring entry. Tested: Tested with Intel AC7260, verified with vmstat -m that new kernel no longer visibly leaks memory from the M_DEVBUF malloc type. Before, leakage was 1KB every few seconds while ping(8)-ing over the wlan connection. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re@ Obtained from: DragonflyBSD.git cc440b26818b5dfdd9af504d71c1b0e6522b53ef Differential Revision: https://reviews.freebsd.org/D6742
2016-06-13 00:13:20 +00:00
struct iwm_rx_data *data = &ring->data[i];
error = bus_dmamap_create(ring->data_dmat, 0, &data->map);
if (error != 0) {
device_printf(sc->sc_dev,
"%s: could not create RX buf DMA map, error %d\n",
__func__, error);
goto fail;
}
data->m = NULL;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if ((error = iwm_rx_addbuf(sc, IWM_RBUF_SIZE, i)) != 0) {
goto fail;
}
}
return 0;
fail: iwm_free_rx_ring(sc, ring);
return error;
}
static void
[iwm] Add bit-polling in Rx-DMA init code path. Taken-From: OpenBSD (if_iwm.c r1.80) Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD ed35558754288911048cb607e57c688273ebd8d4
2016-06-02 05:00:52 +00:00
iwm_disable_rx_dma(struct iwm_softc *sc)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/* XXX conditional nic locks are stupid */
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* XXX print out if we can't lock the NIC? */
if (iwm_nic_lock(sc)) {
/* XXX handle if RX stop doesn't finish? */
(void) iwm_pcie_rx_stop(sc);
iwm_nic_unlock(sc);
}
[iwm] Add bit-polling in Rx-DMA init code path. Taken-From: OpenBSD (if_iwm.c r1.80) Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD ed35558754288911048cb607e57c688273ebd8d4
2016-06-02 05:00:52 +00:00
}
static void
iwm_reset_rx_ring(struct iwm_softc *sc, struct iwm_rx_ring *ring)
{
Ensure the ring state is also blanked upon reset, otherwise duplicate rx events get handled during reset paths. Submitted by: Matthew Dillion <dillon@apollo.backplane.com> Obtained from: DragonflyBSD
2015-09-18 17:39:31 +00:00
/* Reset the ring state */
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ring->cur = 0;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/*
* The hw rx ring index in shared memory must also be cleared,
* otherwise the discrepancy can cause reprocessing chaos.
*/
Ensure the ring state is also blanked upon reset, otherwise duplicate rx events get handled during reset paths. Submitted by: Matthew Dillion <dillon@apollo.backplane.com> Obtained from: DragonflyBSD
2015-09-18 17:39:31 +00:00
memset(sc->rxq.stat, 0, sizeof(*sc->rxq.stat));
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
static void
iwm_free_rx_ring(struct iwm_softc *sc, struct iwm_rx_ring *ring)
{
int i;
iwm_dma_contig_free(&ring->desc_dma);
iwm_dma_contig_free(&ring->stat_dma);
for (i = 0; i < IWM_RX_RING_COUNT; i++) {
struct iwm_rx_data *data = &ring->data[i];
if (data->m != NULL) {
bus_dmamap_sync(ring->data_dmat, data->map,
BUS_DMASYNC_POSTREAD);
bus_dmamap_unload(ring->data_dmat, data->map);
m_freem(data->m);
data->m = NULL;
}
if (data->map != NULL) {
bus_dmamap_destroy(ring->data_dmat, data->map);
data->map = NULL;
}
}
[iwm] Fix up busdma use in the RX path When allocating a new mbuf or bus_dmamap_load()-ing it fails, we can just keep the old mbuf since we are dropping that packet anyway. Instead of doing bus_dmamap_create() and bus_dmamap_destroy() all the time, create an extra bus_dmamap_t which we can use to safely try bus_dmamap_load()-ing the new mbuf. On success we just swap the spare bus_dmamap_t with the data->map of that ring entry. Tested: Tested with Intel AC7260, verified with vmstat -m that new kernel no longer visibly leaks memory from the M_DEVBUF malloc type. Before, leakage was 1KB every few seconds while ping(8)-ing over the wlan connection. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re@ Obtained from: DragonflyBSD.git cc440b26818b5dfdd9af504d71c1b0e6522b53ef Differential Revision: https://reviews.freebsd.org/D6742
2016-06-13 00:13:20 +00:00
if (ring->spare_map != NULL) {
bus_dmamap_destroy(ring->data_dmat, ring->spare_map);
ring->spare_map = NULL;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (ring->data_dmat != NULL) {
bus_dma_tag_destroy(ring->data_dmat);
ring->data_dmat = NULL;
}
}
static int
iwm_alloc_tx_ring(struct iwm_softc *sc, struct iwm_tx_ring *ring, int qid)
{
bus_addr_t paddr;
bus_size_t size;
[iwm] Use mbuf for large firmware commands, like OpenBSD does. We also need to consider the size of large firmware commands in iwm_alloc_tx_ring(), in the dma tag creation, when qid == IWM_MVM_CMD_QUEUE. The old code apparently only allocated a 2KB (MCLBYTES) sized buffer when it actually expected 4KB. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Differential Revision: https://reviews.freebsd.org/D6824
2016-06-23 01:13:30 +00:00
size_t maxsize;
int nsegments;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
int i, error;
ring->qid = qid;
ring->queued = 0;
ring->cur = 0;
/* Allocate TX descriptors (256-byte aligned). */
size = IWM_TX_RING_COUNT * sizeof (struct iwm_tfd);
error = iwm_dma_contig_alloc(sc->sc_dmat, &ring->desc_dma, size, 256);
if (error != 0) {
device_printf(sc->sc_dev,
"could not allocate TX ring DMA memory\n");
goto fail;
}
ring->desc = ring->desc_dma.vaddr;
/*
* We only use rings 0 through 9 (4 EDCA + cmd) so there is no need
* to allocate commands space for other rings.
*/
if (qid > IWM_MVM_CMD_QUEUE)
return 0;
size = IWM_TX_RING_COUNT * sizeof(struct iwm_device_cmd);
error = iwm_dma_contig_alloc(sc->sc_dmat, &ring->cmd_dma, size, 4);
if (error != 0) {
device_printf(sc->sc_dev,
"could not allocate TX cmd DMA memory\n");
goto fail;
}
ring->cmd = ring->cmd_dma.vaddr;
[iwm] Use mbuf for large firmware commands, like OpenBSD does. We also need to consider the size of large firmware commands in iwm_alloc_tx_ring(), in the dma tag creation, when qid == IWM_MVM_CMD_QUEUE. The old code apparently only allocated a 2KB (MCLBYTES) sized buffer when it actually expected 4KB. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Differential Revision: https://reviews.freebsd.org/D6824
2016-06-23 01:13:30 +00:00
/* FW commands may require more mapped space than packets. */
if (qid == IWM_MVM_CMD_QUEUE) {
maxsize = IWM_RBUF_SIZE;
nsegments = 1;
} else {
maxsize = MCLBYTES;
nsegments = IWM_MAX_SCATTER - 2;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
error = bus_dma_tag_create(sc->sc_dmat, 1, 0,
[iwm] Use mbuf for large firmware commands, like OpenBSD does. We also need to consider the size of large firmware commands in iwm_alloc_tx_ring(), in the dma tag creation, when qid == IWM_MVM_CMD_QUEUE. The old code apparently only allocated a 2KB (MCLBYTES) sized buffer when it actually expected 4KB. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Differential Revision: https://reviews.freebsd.org/D6824
2016-06-23 01:13:30 +00:00
BUS_SPACE_MAXADDR_32BIT, BUS_SPACE_MAXADDR, NULL, NULL, maxsize,
nsegments, maxsize, 0, NULL, NULL, &ring->data_dmat);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (error != 0) {
device_printf(sc->sc_dev, "could not create TX buf DMA tag\n");
goto fail;
}
paddr = ring->cmd_dma.paddr;
for (i = 0; i < IWM_TX_RING_COUNT; i++) {
struct iwm_tx_data *data = &ring->data[i];
data->cmd_paddr = paddr;
data->scratch_paddr = paddr + sizeof(struct iwm_cmd_header)
+ offsetof(struct iwm_tx_cmd, scratch);
paddr += sizeof(struct iwm_device_cmd);
error = bus_dmamap_create(ring->data_dmat, 0, &data->map);
if (error != 0) {
device_printf(sc->sc_dev,
"could not create TX buf DMA map\n");
goto fail;
}
}
KASSERT(paddr == ring->cmd_dma.paddr + size,
("invalid physical address"));
return 0;
fail: iwm_free_tx_ring(sc, ring);
return error;
}
static void
iwm_reset_tx_ring(struct iwm_softc *sc, struct iwm_tx_ring *ring)
{
int i;
for (i = 0; i < IWM_TX_RING_COUNT; i++) {
struct iwm_tx_data *data = &ring->data[i];
if (data->m != NULL) {
bus_dmamap_sync(ring->data_dmat, data->map,
BUS_DMASYNC_POSTWRITE);
bus_dmamap_unload(ring->data_dmat, data->map);
m_freem(data->m);
data->m = NULL;
}
}
/* Clear TX descriptors. */
memset(ring->desc, 0, ring->desc_dma.size);
bus_dmamap_sync(ring->desc_dma.tag, ring->desc_dma.map,
BUS_DMASYNC_PREWRITE);
sc->qfullmsk &= ~(1 << ring->qid);
ring->queued = 0;
ring->cur = 0;
}
static void
iwm_free_tx_ring(struct iwm_softc *sc, struct iwm_tx_ring *ring)
{
int i;
iwm_dma_contig_free(&ring->desc_dma);
iwm_dma_contig_free(&ring->cmd_dma);
for (i = 0; i < IWM_TX_RING_COUNT; i++) {
struct iwm_tx_data *data = &ring->data[i];
if (data->m != NULL) {
bus_dmamap_sync(ring->data_dmat, data->map,
BUS_DMASYNC_POSTWRITE);
bus_dmamap_unload(ring->data_dmat, data->map);
m_freem(data->m);
data->m = NULL;
}
if (data->map != NULL) {
bus_dmamap_destroy(ring->data_dmat, data->map);
data->map = NULL;
}
}
if (ring->data_dmat != NULL) {
bus_dma_tag_destroy(ring->data_dmat);
ring->data_dmat = NULL;
}
}
/*
* High-level hardware frobbing routines
*/
static void
iwm_enable_interrupts(struct iwm_softc *sc)
{
sc->sc_intmask = IWM_CSR_INI_SET_MASK;
IWM_WRITE(sc, IWM_CSR_INT_MASK, sc->sc_intmask);
}
static void
iwm_restore_interrupts(struct iwm_softc *sc)
{
IWM_WRITE(sc, IWM_CSR_INT_MASK, sc->sc_intmask);
}
static void
iwm_disable_interrupts(struct iwm_softc *sc)
{
/* disable interrupts */
IWM_WRITE(sc, IWM_CSR_INT_MASK, 0);
/* acknowledge all interrupts */
IWM_WRITE(sc, IWM_CSR_INT, ~0);
IWM_WRITE(sc, IWM_CSR_FH_INT_STATUS, ~0);
}
static void
iwm_ict_reset(struct iwm_softc *sc)
{
iwm_disable_interrupts(sc);
/* Reset ICT table. */
memset(sc->ict_dma.vaddr, 0, IWM_ICT_SIZE);
sc->ict_cur = 0;
/* Set physical address of ICT table (4KB aligned). */
IWM_WRITE(sc, IWM_CSR_DRAM_INT_TBL_REG,
IWM_CSR_DRAM_INT_TBL_ENABLE
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
| IWM_CSR_DRAM_INIT_TBL_WRITE_POINTER
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
| IWM_CSR_DRAM_INIT_TBL_WRAP_CHECK
| sc->ict_dma.paddr >> IWM_ICT_PADDR_SHIFT);
/* Switch to ICT interrupt mode in driver. */
sc->sc_flags |= IWM_FLAG_USE_ICT;
/* Re-enable interrupts. */
IWM_WRITE(sc, IWM_CSR_INT, ~0);
iwm_enable_interrupts(sc);
}
/* iwlwifi pcie/trans.c */
/*
* Since this .. hard-resets things, it's time to actually
* mark the first vap (if any) as having no mac context.
* It's annoying, but since the driver is potentially being
* stop/start'ed whilst active (thanks openbsd port!) we
* have to correctly track this.
*/
static void
iwm_stop_device(struct iwm_softc *sc)
{
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct ieee80211com *ic = &sc->sc_ic;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
struct ieee80211vap *vap = TAILQ_FIRST(&ic->ic_vaps);
[iwm] When stopping TX DMA, wait for all channels at once. * Makes the TX DMA stopping more similar to Linux code, and potentially a bit faster. Also, output an error message when TX DMA idling fails. Taken-From: Linux iwlwifi Tested: * AC3165, STA mode Approved by: adrian (mentor) Obtained from: DragonFlyBSD git 2ee486ddff973ac552ff787c17e8d83e8ae0f24c Differential Revision: https://reviews.freebsd.org/D7325
2016-07-27 20:51:31 +00:00
int chnl, qid;
uint32_t mask = 0;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* tell the device to stop sending interrupts */
iwm_disable_interrupts(sc);
/*
* FreeBSD-local: mark the first vap as not-uploaded,
* so the next transition through auth/assoc
* will correctly populate the MAC context.
*/
if (vap) {
struct iwm_vap *iv = IWM_VAP(vap);
iv->is_uploaded = 0;
}
/* device going down, Stop using ICT table */
sc->sc_flags &= ~IWM_FLAG_USE_ICT;
/* stop tx and rx. tx and rx bits, as usual, are from if_iwn */
iwm_write_prph(sc, IWM_SCD_TXFACT, 0);
if (iwm_nic_lock(sc)) {
[iwm] When stopping TX DMA, wait for all channels at once. * Makes the TX DMA stopping more similar to Linux code, and potentially a bit faster. Also, output an error message when TX DMA idling fails. Taken-From: Linux iwlwifi Tested: * AC3165, STA mode Approved by: adrian (mentor) Obtained from: DragonFlyBSD git 2ee486ddff973ac552ff787c17e8d83e8ae0f24c Differential Revision: https://reviews.freebsd.org/D7325
2016-07-27 20:51:31 +00:00
/* Stop each Tx DMA channel */
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
for (chnl = 0; chnl < IWM_FH_TCSR_CHNL_NUM; chnl++) {
IWM_WRITE(sc,
IWM_FH_TCSR_CHNL_TX_CONFIG_REG(chnl), 0);
[iwm] When stopping TX DMA, wait for all channels at once. * Makes the TX DMA stopping more similar to Linux code, and potentially a bit faster. Also, output an error message when TX DMA idling fails. Taken-From: Linux iwlwifi Tested: * AC3165, STA mode Approved by: adrian (mentor) Obtained from: DragonFlyBSD git 2ee486ddff973ac552ff787c17e8d83e8ae0f24c Differential Revision: https://reviews.freebsd.org/D7325
2016-07-27 20:51:31 +00:00
mask |= IWM_FH_TSSR_TX_STATUS_REG_MSK_CHNL_IDLE(chnl);
}
/* Wait for DMA channels to be idle */
[iwm] Fix iwm_poll_bit() usage in iwm_stop_device(), fixup r303418. * iwm_poll_bit() returns 1 on success and 0 on failure, whereas iwl_poll_bit() in Linux's iwlwifi returns >= 0 on success and < 0 on failure. * Because of the wrong iwm_poll_bit return code check, no warning was printed if tx DMA stopping failed. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7371
2016-07-30 19:03:32 +00:00
if (!iwm_poll_bit(sc, IWM_FH_TSSR_TX_STATUS_REG, mask, mask,
5000)) {
[iwm] When stopping TX DMA, wait for all channels at once. * Makes the TX DMA stopping more similar to Linux code, and potentially a bit faster. Also, output an error message when TX DMA idling fails. Taken-From: Linux iwlwifi Tested: * AC3165, STA mode Approved by: adrian (mentor) Obtained from: DragonFlyBSD git 2ee486ddff973ac552ff787c17e8d83e8ae0f24c Differential Revision: https://reviews.freebsd.org/D7325
2016-07-27 20:51:31 +00:00
device_printf(sc->sc_dev,
"Failing on timeout while stopping DMA channel: [0x%08x]\n",
IWM_READ(sc, IWM_FH_TSSR_TX_STATUS_REG));
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
iwm_nic_unlock(sc);
}
[iwm] Add bit-polling in Rx-DMA init code path. Taken-From: OpenBSD (if_iwm.c r1.80) Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD ed35558754288911048cb607e57c688273ebd8d4
2016-06-02 05:00:52 +00:00
iwm_disable_rx_dma(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Stop RX ring. */
iwm_reset_rx_ring(sc, &sc->rxq);
/* Reset all TX rings. */
for (qid = 0; qid < nitems(sc->txq); qid++)
iwm_reset_tx_ring(sc, &sc->txq[qid]);
/*
* Power-down device's busmaster DMA clocks
*/
iwm_write_prph(sc, IWM_APMG_CLK_DIS_REG, IWM_APMG_CLK_VAL_DMA_CLK_RQT);
DELAY(5);
/* Make sure (redundant) we've released our request to stay awake */
IWM_CLRBITS(sc, IWM_CSR_GP_CNTRL,
IWM_CSR_GP_CNTRL_REG_FLAG_MAC_ACCESS_REQ);
/* Stop the device, and put it in low power state */
iwm_apm_stop(sc);
/* Upon stop, the APM issues an interrupt if HW RF kill is set.
* Clean again the interrupt here
*/
iwm_disable_interrupts(sc);
/* stop and reset the on-board processor */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
IWM_WRITE(sc, IWM_CSR_RESET, IWM_CSR_RESET_REG_FLAG_SW_RESET);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* Even if we stop the HW, we still want the RF kill
* interrupt
*/
iwm_enable_rfkill_int(sc);
iwm_check_rfkill(sc);
}
/* iwlwifi: mvm/ops.c */
static void
iwm_mvm_nic_config(struct iwm_softc *sc)
{
uint8_t radio_cfg_type, radio_cfg_step, radio_cfg_dash;
uint32_t reg_val = 0;
radio_cfg_type = (sc->sc_fw_phy_config & IWM_FW_PHY_CFG_RADIO_TYPE) >>
IWM_FW_PHY_CFG_RADIO_TYPE_POS;
radio_cfg_step = (sc->sc_fw_phy_config & IWM_FW_PHY_CFG_RADIO_STEP) >>
IWM_FW_PHY_CFG_RADIO_STEP_POS;
radio_cfg_dash = (sc->sc_fw_phy_config & IWM_FW_PHY_CFG_RADIO_DASH) >>
IWM_FW_PHY_CFG_RADIO_DASH_POS;
/* SKU control */
reg_val |= IWM_CSR_HW_REV_STEP(sc->sc_hw_rev) <<
IWM_CSR_HW_IF_CONFIG_REG_POS_MAC_STEP;
reg_val |= IWM_CSR_HW_REV_DASH(sc->sc_hw_rev) <<
IWM_CSR_HW_IF_CONFIG_REG_POS_MAC_DASH;
/* radio configuration */
reg_val |= radio_cfg_type << IWM_CSR_HW_IF_CONFIG_REG_POS_PHY_TYPE;
reg_val |= radio_cfg_step << IWM_CSR_HW_IF_CONFIG_REG_POS_PHY_STEP;
reg_val |= radio_cfg_dash << IWM_CSR_HW_IF_CONFIG_REG_POS_PHY_DASH;
IWM_WRITE(sc, IWM_CSR_HW_IF_CONFIG_REG, reg_val);
IWM_DPRINTF(sc, IWM_DEBUG_RESET,
"Radio type=0x%x-0x%x-0x%x\n", radio_cfg_type,
radio_cfg_step, radio_cfg_dash);
/*
* W/A : NIC is stuck in a reset state after Early PCIe power off
* (PCIe power is lost before PERST# is asserted), causing ME FW
* to lose ownership and not being able to obtain it back.
*/
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (sc->sc_device_family == IWM_DEVICE_FAMILY_7000) {
iwm_set_bits_mask_prph(sc, IWM_APMG_PS_CTRL_REG,
IWM_APMG_PS_CTRL_EARLY_PWR_OFF_RESET_DIS,
~IWM_APMG_PS_CTRL_EARLY_PWR_OFF_RESET_DIS);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
static int
iwm_nic_rx_init(struct iwm_softc *sc)
{
if (!iwm_nic_lock(sc))
return EBUSY;
/*
* Initialize RX ring. This is from the iwn driver.
*/
memset(sc->rxq.stat, 0, sizeof(*sc->rxq.stat));
/* stop DMA */
[iwm] Add bit-polling in Rx-DMA init code path. Taken-From: OpenBSD (if_iwm.c r1.80) Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD ed35558754288911048cb607e57c688273ebd8d4
2016-06-02 05:00:52 +00:00
iwm_disable_rx_dma(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_WRITE(sc, IWM_FH_MEM_RCSR_CHNL0_RBDCB_WPTR, 0);
IWM_WRITE(sc, IWM_FH_MEM_RCSR_CHNL0_FLUSH_RB_REQ, 0);
IWM_WRITE(sc, IWM_FH_RSCSR_CHNL0_RDPTR, 0);
IWM_WRITE(sc, IWM_FH_RSCSR_CHNL0_RBDCB_WPTR_REG, 0);
/* Set physical address of RX ring (256-byte aligned). */
IWM_WRITE(sc,
IWM_FH_RSCSR_CHNL0_RBDCB_BASE_REG, sc->rxq.desc_dma.paddr >> 8);
/* Set physical address of RX status (16-byte aligned). */
IWM_WRITE(sc,
IWM_FH_RSCSR_CHNL0_STTS_WPTR_REG, sc->rxq.stat_dma.paddr >> 4);
/* Enable RX. */
IWM_WRITE(sc, IWM_FH_MEM_RCSR_CHNL0_CONFIG_REG,
IWM_FH_RCSR_RX_CONFIG_CHNL_EN_ENABLE_VAL |
IWM_FH_RCSR_CHNL0_RX_IGNORE_RXF_EMPTY | /* HW bug */
IWM_FH_RCSR_CHNL0_RX_CONFIG_IRQ_DEST_INT_HOST_VAL |
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
IWM_FH_RCSR_CHNL0_RX_CONFIG_SINGLE_FRAME_MSK |
(IWM_RX_RB_TIMEOUT << IWM_FH_RCSR_RX_CONFIG_REG_IRQ_RBTH_POS) |
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_FH_RCSR_RX_CONFIG_REG_VAL_RB_SIZE_4K |
IWM_RX_QUEUE_SIZE_LOG << IWM_FH_RCSR_RX_CONFIG_RBDCB_SIZE_POS);
IWM_WRITE_1(sc, IWM_CSR_INT_COALESCING, IWM_HOST_INT_TIMEOUT_DEF);
/* W/A for interrupt coalescing bug in 7260 and 3160 */
if (sc->host_interrupt_operation_mode)
IWM_SETBITS(sc, IWM_CSR_INT_COALESCING, IWM_HOST_INT_OPER_MODE);
/*
* Thus sayeth el jefe (iwlwifi) via a comment:
*
* This value should initially be 0 (before preparing any
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
* RBs), should be 8 after preparing the first 8 RBs (for example)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
*/
IWM_WRITE(sc, IWM_FH_RSCSR_CHNL0_WPTR, 8);
iwm_nic_unlock(sc);
return 0;
}
static int
iwm_nic_tx_init(struct iwm_softc *sc)
{
int qid;
if (!iwm_nic_lock(sc))
return EBUSY;
/* Deactivate TX scheduler. */
iwm_write_prph(sc, IWM_SCD_TXFACT, 0);
/* Set physical address of "keep warm" page (16-byte aligned). */
IWM_WRITE(sc, IWM_FH_KW_MEM_ADDR_REG, sc->kw_dma.paddr >> 4);
/* Initialize TX rings. */
for (qid = 0; qid < nitems(sc->txq); qid++) {
struct iwm_tx_ring *txq = &sc->txq[qid];
/* Set physical address of TX ring (256-byte aligned). */
IWM_WRITE(sc, IWM_FH_MEM_CBBC_QUEUE(qid),
txq->desc_dma.paddr >> 8);
IWM_DPRINTF(sc, IWM_DEBUG_XMIT,
"%s: loading ring %d descriptors (%p) at %lx\n",
__func__,
qid, txq->desc,
(unsigned long) (txq->desc_dma.paddr >> 8));
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
iwm_write_prph(sc, IWM_SCD_GP_CTRL, IWM_SCD_GP_CTRL_AUTO_ACTIVE_MODE);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_nic_unlock(sc);
return 0;
}
static int
iwm_nic_init(struct iwm_softc *sc)
{
int error;
iwm_apm_init(sc);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (sc->sc_device_family == IWM_DEVICE_FAMILY_7000)
iwm_set_pwr(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_mvm_nic_config(sc);
if ((error = iwm_nic_rx_init(sc)) != 0)
return error;
/*
* Ditto for TX, from iwn
*/
if ((error = iwm_nic_tx_init(sc)) != 0)
return error;
IWM_DPRINTF(sc, IWM_DEBUG_RESET,
"%s: shadow registers enabled\n", __func__);
IWM_SETBITS(sc, IWM_CSR_MAC_SHADOW_REG_CTRL, 0x800fffff);
return 0;
}
const uint8_t iwm_mvm_ac_to_tx_fifo[] = {
IWM_MVM_TX_FIFO_VO,
IWM_MVM_TX_FIFO_VI,
IWM_MVM_TX_FIFO_BE,
IWM_MVM_TX_FIFO_BK,
};
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
static int
iwm_enable_txq(struct iwm_softc *sc, int sta_id, int qid, int fifo)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
if (!iwm_nic_lock(sc)) {
device_printf(sc->sc_dev,
"%s: cannot enable txq %d\n",
__func__,
qid);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
return EBUSY;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
IWM_WRITE(sc, IWM_HBUS_TARG_WRPTR, qid << 8 | 0);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (qid == IWM_MVM_CMD_QUEUE) {
/* unactivate before configuration */
iwm_write_prph(sc, IWM_SCD_QUEUE_STATUS_BITS(qid),
(0 << IWM_SCD_QUEUE_STTS_REG_POS_ACTIVE)
| (1 << IWM_SCD_QUEUE_STTS_REG_POS_SCD_ACT_EN));
iwm_clear_bits_prph(sc, IWM_SCD_AGGR_SEL, (1 << qid));
iwm_write_prph(sc, IWM_SCD_QUEUE_RDPTR(qid), 0);
iwm_write_mem32(sc, sc->sched_base + IWM_SCD_CONTEXT_QUEUE_OFFSET(qid), 0);
/* Set scheduler window size and frame limit. */
iwm_write_mem32(sc,
sc->sched_base + IWM_SCD_CONTEXT_QUEUE_OFFSET(qid) +
sizeof(uint32_t),
((IWM_FRAME_LIMIT << IWM_SCD_QUEUE_CTX_REG2_WIN_SIZE_POS) &
IWM_SCD_QUEUE_CTX_REG2_WIN_SIZE_MSK) |
((IWM_FRAME_LIMIT << IWM_SCD_QUEUE_CTX_REG2_FRAME_LIMIT_POS) &
IWM_SCD_QUEUE_CTX_REG2_FRAME_LIMIT_MSK));
iwm_write_prph(sc, IWM_SCD_QUEUE_STATUS_BITS(qid),
(1 << IWM_SCD_QUEUE_STTS_REG_POS_ACTIVE) |
(fifo << IWM_SCD_QUEUE_STTS_REG_POS_TXF) |
(1 << IWM_SCD_QUEUE_STTS_REG_POS_WSL) |
IWM_SCD_QUEUE_STTS_REG_MSK);
} else {
struct iwm_scd_txq_cfg_cmd cmd;
int error;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
iwm_nic_unlock(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
memset(&cmd, 0, sizeof(cmd));
cmd.scd_queue = qid;
cmd.enable = 1;
cmd.sta_id = sta_id;
cmd.tx_fifo = fifo;
cmd.aggregate = 0;
cmd.window = IWM_FRAME_LIMIT;
error = iwm_mvm_send_cmd_pdu(sc, IWM_SCD_QUEUE_CFG, IWM_CMD_SYNC,
sizeof(cmd), &cmd);
if (error) {
device_printf(sc->sc_dev,
"cannot enable txq %d\n", qid);
return error;
}
if (!iwm_nic_lock(sc))
return EBUSY;
}
iwm_write_prph(sc, IWM_SCD_EN_CTRL,
iwm_read_prph(sc, IWM_SCD_EN_CTRL) | qid);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_nic_unlock(sc);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_XMIT, "%s: enabled txq %d FIFO %d\n",
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
__func__, qid, fifo);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
return 0;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
static int
iwm_post_alive(struct iwm_softc *sc)
{
int nwords;
int error, chnl;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
uint32_t base;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (!iwm_nic_lock(sc))
return EBUSY;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
base = iwm_read_prph(sc, IWM_SCD_SRAM_BASE_ADDR);
if (sc->sched_base != base) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
"%s: sched addr mismatch: alive: 0x%x prph: 0x%x\n",
__func__, sc->sched_base, base);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
iwm_ict_reset(sc);
/* Clear TX scheduler state in SRAM. */
nwords = (IWM_SCD_TRANS_TBL_MEM_UPPER_BOUND -
IWM_SCD_CONTEXT_MEM_LOWER_BOUND)
/ sizeof(uint32_t);
error = iwm_write_mem(sc,
sc->sched_base + IWM_SCD_CONTEXT_MEM_LOWER_BOUND,
NULL, nwords);
if (error)
goto out;
/* Set physical address of TX scheduler rings (1KB aligned). */
iwm_write_prph(sc, IWM_SCD_DRAM_BASE_ADDR, sc->sched_dma.paddr >> 10);
iwm_write_prph(sc, IWM_SCD_CHAINEXT_EN, 0);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
iwm_nic_unlock(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* enable command channel */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
error = iwm_enable_txq(sc, 0 /* unused */, IWM_MVM_CMD_QUEUE, 7);
if (error)
return error;
if (!iwm_nic_lock(sc))
return EBUSY;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_write_prph(sc, IWM_SCD_TXFACT, 0xff);
/* Enable DMA channels. */
for (chnl = 0; chnl < IWM_FH_TCSR_CHNL_NUM; chnl++) {
IWM_WRITE(sc, IWM_FH_TCSR_CHNL_TX_CONFIG_REG(chnl),
IWM_FH_TCSR_TX_CONFIG_REG_VAL_DMA_CHNL_ENABLE |
IWM_FH_TCSR_TX_CONFIG_REG_VAL_DMA_CREDIT_ENABLE);
}
IWM_SETBITS(sc, IWM_FH_TX_CHICKEN_BITS_REG,
IWM_FH_TX_CHICKEN_BITS_SCD_AUTO_RETRY_EN);
/* Enable L1-Active */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (sc->sc_device_family != IWM_DEVICE_FAMILY_8000) {
iwm_clear_bits_prph(sc, IWM_APMG_PCIDEV_STT_REG,
IWM_APMG_PCIDEV_STT_VAL_L1_ACT_DIS);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
out:
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
iwm_nic_unlock(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
}
/*
* NVM read access and content parsing. We do not support
* external NVM or writing NVM.
* iwlwifi/mvm/nvm.c
*/
/* list of NVM sections we are allowed/need to read */
const int nvm_to_read[] = {
IWM_NVM_SECTION_TYPE_HW,
IWM_NVM_SECTION_TYPE_SW,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
IWM_NVM_SECTION_TYPE_REGULATORY,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_NVM_SECTION_TYPE_CALIBRATION,
IWM_NVM_SECTION_TYPE_PRODUCTION,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
IWM_NVM_SECTION_TYPE_HW_8000,
IWM_NVM_SECTION_TYPE_MAC_OVERRIDE,
IWM_NVM_SECTION_TYPE_PHY_SKU,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
};
/* Default NVM size to read */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
#define IWM_NVM_DEFAULT_CHUNK_SIZE (2*1024)
#define IWM_MAX_NVM_SECTION_SIZE 8192
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
#define IWM_NVM_WRITE_OPCODE 1
#define IWM_NVM_READ_OPCODE 0
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/* load nvm chunk response */
[iwm] Sync iwm_nvm_read_chunk() function with Linux iwlwifi. This fixes an error handling detail in iwm_nvm_read_chunk(), where an error response from the firmware for an NVM read shouldn't be fatal if the offset was non-zero. Approved by: adrian (mentor) Obtained from: DragonFlyBSD git 250a1c33fca1725121fe499f9cebc90267d209f9 Differential Revision: https://reviews.freebsd.org/D8542
2016-11-17 20:00:20 +00:00
enum {
IWM_READ_NVM_CHUNK_SUCCEED = 0,
IWM_READ_NVM_CHUNK_NOT_VALID_ADDRESS = 1
};
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int
iwm_nvm_read_chunk(struct iwm_softc *sc, uint16_t section,
uint16_t offset, uint16_t length, uint8_t *data, uint16_t *len)
{
offset = 0;
struct iwm_nvm_access_cmd nvm_access_cmd = {
.offset = htole16(offset),
.length = htole16(length),
.type = htole16(section),
.op_code = IWM_NVM_READ_OPCODE,
};
struct iwm_nvm_access_resp *nvm_resp;
struct iwm_rx_packet *pkt;
struct iwm_host_cmd cmd = {
.id = IWM_NVM_ACCESS_CMD,
[iwm] Sync iwm_nvm_read_chunk() function with Linux iwlwifi. This fixes an error handling detail in iwm_nvm_read_chunk(), where an error response from the firmware for an NVM read shouldn't be fatal if the offset was non-zero. Approved by: adrian (mentor) Obtained from: DragonFlyBSD git 250a1c33fca1725121fe499f9cebc90267d209f9 Differential Revision: https://reviews.freebsd.org/D8542
2016-11-17 20:00:20 +00:00
.flags = IWM_CMD_WANT_SKB | IWM_CMD_SEND_IN_RFKILL,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
.data = { &nvm_access_cmd, },
};
[iwm] Sync iwm_nvm_read_chunk() function with Linux iwlwifi. This fixes an error handling detail in iwm_nvm_read_chunk(), where an error response from the firmware for an NVM read shouldn't be fatal if the offset was non-zero. Approved by: adrian (mentor) Obtained from: DragonFlyBSD git 250a1c33fca1725121fe499f9cebc90267d209f9 Differential Revision: https://reviews.freebsd.org/D8542
2016-11-17 20:00:20 +00:00
int ret, bytes_read, offset_read;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
uint8_t *resp_data;
cmd.len[0] = sizeof(struct iwm_nvm_access_cmd);
ret = iwm_send_cmd(sc, &cmd);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (ret) {
device_printf(sc->sc_dev,
"Could not send NVM_ACCESS command (error=%d)\n", ret);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return ret;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
pkt = cmd.resp_pkt;
if (pkt->hdr.flags & IWM_CMD_FAILED_MSK) {
device_printf(sc->sc_dev,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
"Bad return from IWM_NVM_ACCES_COMMAND (0x%08X)\n",
pkt->hdr.flags);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ret = EIO;
goto exit;
}
/* Extract NVM response */
nvm_resp = (void *)pkt->data;
ret = le16toh(nvm_resp->status);
bytes_read = le16toh(nvm_resp->length);
offset_read = le16toh(nvm_resp->offset);
resp_data = nvm_resp->data;
if (ret) {
[iwm] Sync iwm_nvm_read_chunk() function with Linux iwlwifi. This fixes an error handling detail in iwm_nvm_read_chunk(), where an error response from the firmware for an NVM read shouldn't be fatal if the offset was non-zero. Approved by: adrian (mentor) Obtained from: DragonFlyBSD git 250a1c33fca1725121fe499f9cebc90267d209f9 Differential Revision: https://reviews.freebsd.org/D8542
2016-11-17 20:00:20 +00:00
if ((offset != 0) &&
(ret == IWM_READ_NVM_CHUNK_NOT_VALID_ADDRESS)) {
/*
* meaning of NOT_VALID_ADDRESS:
* driver try to read chunk from address that is
* multiple of 2K and got an error since addr is empty.
* meaning of (offset != 0): driver already
* read valid data from another chunk so this case
* is not an error.
*/
IWM_DPRINTF(sc, IWM_DEBUG_EEPROM | IWM_DEBUG_RESET,
"NVM access command failed on offset 0x%x since that section size is multiple 2K\n",
offset);
*len = 0;
ret = 0;
} else {
IWM_DPRINTF(sc, IWM_DEBUG_EEPROM | IWM_DEBUG_RESET,
"NVM access command failed with status %d\n", ret);
ret = EIO;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
goto exit;
}
if (offset_read != offset) {
device_printf(sc->sc_dev,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
"NVM ACCESS response with invalid offset %d\n",
offset_read);
ret = EINVAL;
goto exit;
}
if (bytes_read > length) {
device_printf(sc->sc_dev,
"NVM ACCESS response with too much data "
[iwm] Sync iwm_nvm_read_chunk() function with Linux iwlwifi. This fixes an error handling detail in iwm_nvm_read_chunk(), where an error response from the firmware for an NVM read shouldn't be fatal if the offset was non-zero. Approved by: adrian (mentor) Obtained from: DragonFlyBSD git 250a1c33fca1725121fe499f9cebc90267d209f9 Differential Revision: https://reviews.freebsd.org/D8542
2016-11-17 20:00:20 +00:00
"(%d bytes requested, %d bytes received)\n",
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
length, bytes_read);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ret = EINVAL;
goto exit;
}
memcpy(data + offset, resp_data, bytes_read);
*len = bytes_read;
exit:
iwm_free_resp(sc, &cmd);
return ret;
}
/*
* Reads an NVM section completely.
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
* NICs prior to 7000 family don't have a real NVM, but just read
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
* section 0 which is the EEPROM. Because the EEPROM reading is unlimited
* by uCode, we need to manually check in this case that we don't
* overflow and try to read more than the EEPROM size.
* For 7000 family NICs, we supply the maximal size we can read, and
* the uCode fills the response with as much data as we can,
* without overflowing, so no check is needed.
*/
static int
iwm_nvm_read_section(struct iwm_softc *sc,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
uint16_t section, uint8_t *data, uint16_t *len, size_t max_len)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
uint16_t chunklen, seglen;
int error = 0;
IWM_DPRINTF(sc, IWM_DEBUG_RESET,
"reading NVM section %d\n", section);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
chunklen = seglen = IWM_NVM_DEFAULT_CHUNK_SIZE;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
*len = 0;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/* Read NVM chunks until exhausted (reading less than requested) */
while (seglen == chunklen && *len < max_len) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
error = iwm_nvm_read_chunk(sc,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
section, *len, chunklen, data, &seglen);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (error) {
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_RESET,
"Cannot read from NVM section "
"%d at offset %d\n", section, *len);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
}
*len += seglen;
}
IWM_DPRINTF(sc, IWM_DEBUG_RESET,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
"NVM section %d read completed (%d bytes, error=%d)\n",
section, *len, error);
return error;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
/*
* BEGIN IWM_NVM_PARSE
*/
/* iwlwifi/iwl-nvm-parse.c */
/* NVM offsets (in words) definitions */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
enum iwm_nvm_offsets {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* NVM HW-Section offset (in words) definitions */
IWM_HW_ADDR = 0x15,
/* NVM SW-Section offset (in words) definitions */
IWM_NVM_SW_SECTION = 0x1C0,
IWM_NVM_VERSION = 0,
IWM_RADIO_CFG = 1,
IWM_SKU = 2,
IWM_N_HW_ADDRS = 3,
IWM_NVM_CHANNELS = 0x1E0 - IWM_NVM_SW_SECTION,
/* NVM calibration section offset (in words) definitions */
IWM_NVM_CALIB_SECTION = 0x2B8,
IWM_XTAL_CALIB = 0x316 - IWM_NVM_CALIB_SECTION
};
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
enum iwm_8000_nvm_offsets {
/* NVM HW-Section offset (in words) definitions */
IWM_HW_ADDR0_WFPM_8000 = 0x12,
IWM_HW_ADDR1_WFPM_8000 = 0x16,
IWM_HW_ADDR0_PCIE_8000 = 0x8A,
IWM_HW_ADDR1_PCIE_8000 = 0x8E,
IWM_MAC_ADDRESS_OVERRIDE_8000 = 1,
/* NVM SW-Section offset (in words) definitions */
IWM_NVM_SW_SECTION_8000 = 0x1C0,
IWM_NVM_VERSION_8000 = 0,
IWM_RADIO_CFG_8000 = 0,
IWM_SKU_8000 = 2,
IWM_N_HW_ADDRS_8000 = 3,
/* NVM REGULATORY -Section offset (in words) definitions */
IWM_NVM_CHANNELS_8000 = 0,
IWM_NVM_LAR_OFFSET_8000_OLD = 0x4C7,
IWM_NVM_LAR_OFFSET_8000 = 0x507,
IWM_NVM_LAR_ENABLED_8000 = 0x7,
/* NVM calibration section offset (in words) definitions */
IWM_NVM_CALIB_SECTION_8000 = 0x2B8,
IWM_XTAL_CALIB_8000 = 0x316 - IWM_NVM_CALIB_SECTION_8000
};
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* SKU Capabilities (actual values from NVM definition) */
enum nvm_sku_bits {
IWM_NVM_SKU_CAP_BAND_24GHZ = (1 << 0),
IWM_NVM_SKU_CAP_BAND_52GHZ = (1 << 1),
IWM_NVM_SKU_CAP_11N_ENABLE = (1 << 2),
IWM_NVM_SKU_CAP_11AC_ENABLE = (1 << 3),
};
/* radio config bits (actual values from NVM definition) */
#define IWM_NVM_RF_CFG_DASH_MSK(x) (x & 0x3) /* bits 0-1 */
#define IWM_NVM_RF_CFG_STEP_MSK(x) ((x >> 2) & 0x3) /* bits 2-3 */
#define IWM_NVM_RF_CFG_TYPE_MSK(x) ((x >> 4) & 0x3) /* bits 4-5 */
#define IWM_NVM_RF_CFG_PNUM_MSK(x) ((x >> 6) & 0x3) /* bits 6-7 */
#define IWM_NVM_RF_CFG_TX_ANT_MSK(x) ((x >> 8) & 0xF) /* bits 8-11 */
#define IWM_NVM_RF_CFG_RX_ANT_MSK(x) ((x >> 12) & 0xF) /* bits 12-15 */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
#define IWM_NVM_RF_CFG_FLAVOR_MSK_8000(x) (x & 0xF)
#define IWM_NVM_RF_CFG_DASH_MSK_8000(x) ((x >> 4) & 0xF)
#define IWM_NVM_RF_CFG_STEP_MSK_8000(x) ((x >> 8) & 0xF)
#define IWM_NVM_RF_CFG_TYPE_MSK_8000(x) ((x >> 12) & 0xFFF)
#define IWM_NVM_RF_CFG_TX_ANT_MSK_8000(x) ((x >> 24) & 0xF)
#define IWM_NVM_RF_CFG_RX_ANT_MSK_8000(x) ((x >> 28) & 0xF)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
#define DEFAULT_MAX_TX_POWER 16
/**
* enum iwm_nvm_channel_flags - channel flags in NVM
* @IWM_NVM_CHANNEL_VALID: channel is usable for this SKU/geo
* @IWM_NVM_CHANNEL_IBSS: usable as an IBSS channel
* @IWM_NVM_CHANNEL_ACTIVE: active scanning allowed
* @IWM_NVM_CHANNEL_RADAR: radar detection required
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
* XXX cannot find this (DFS) flag in iwl-nvm-parse.c
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
* @IWM_NVM_CHANNEL_DFS: dynamic freq selection candidate
* @IWM_NVM_CHANNEL_WIDE: 20 MHz channel okay (?)
* @IWM_NVM_CHANNEL_40MHZ: 40 MHz channel okay (?)
* @IWM_NVM_CHANNEL_80MHZ: 80 MHz channel okay (?)
* @IWM_NVM_CHANNEL_160MHZ: 160 MHz channel okay (?)
*/
enum iwm_nvm_channel_flags {
IWM_NVM_CHANNEL_VALID = (1 << 0),
IWM_NVM_CHANNEL_IBSS = (1 << 1),
IWM_NVM_CHANNEL_ACTIVE = (1 << 3),
IWM_NVM_CHANNEL_RADAR = (1 << 4),
IWM_NVM_CHANNEL_DFS = (1 << 7),
IWM_NVM_CHANNEL_WIDE = (1 << 8),
IWM_NVM_CHANNEL_40MHZ = (1 << 9),
IWM_NVM_CHANNEL_80MHZ = (1 << 10),
IWM_NVM_CHANNEL_160MHZ = (1 << 11),
};
/*
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
* Translate EEPROM flags to net80211.
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
*/
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
static uint32_t
iwm_eeprom_channel_flags(uint16_t ch_flags)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
uint32_t nflags;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
nflags = 0;
if ((ch_flags & IWM_NVM_CHANNEL_ACTIVE) == 0)
nflags |= IEEE80211_CHAN_PASSIVE;
if ((ch_flags & IWM_NVM_CHANNEL_IBSS) == 0)
nflags |= IEEE80211_CHAN_NOADHOC;
if (ch_flags & IWM_NVM_CHANNEL_RADAR) {
nflags |= IEEE80211_CHAN_DFS;
/* Just in case. */
nflags |= IEEE80211_CHAN_NOADHOC;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
return (nflags);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
static void
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
iwm_add_channel_band(struct iwm_softc *sc, struct ieee80211_channel chans[],
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
int maxchans, int *nchans, int ch_idx, size_t ch_num,
const uint8_t bands[])
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
const uint16_t * const nvm_ch_flags = sc->sc_nvm.nvm_ch_flags;
uint32_t nflags;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
uint16_t ch_flags;
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
uint8_t ieee;
int error;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
for (; ch_idx < ch_num; ch_idx++) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ch_flags = le16_to_cpup(nvm_ch_flags + ch_idx);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (sc->sc_device_family == IWM_DEVICE_FAMILY_7000)
ieee = iwm_nvm_channels[ch_idx];
else
ieee = iwm_nvm_channels_8000[ch_idx];
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (!(ch_flags & IWM_NVM_CHANNEL_VALID)) {
IWM_DPRINTF(sc, IWM_DEBUG_EEPROM,
"Ch. %d Flags %x [%sGHz] - No traffic\n",
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
ieee, ch_flags,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
(ch_idx >= IWM_NUM_2GHZ_CHANNELS) ?
"5.2" : "2.4");
continue;
}
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
nflags = iwm_eeprom_channel_flags(ch_flags);
error = ieee80211_add_channel(chans, maxchans, nchans,
ieee, 0, 0, nflags, bands);
if (error != 0)
break;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_EEPROM,
"Ch. %d Flags %x [%sGHz] - Added\n",
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
ieee, ch_flags,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
(ch_idx >= IWM_NUM_2GHZ_CHANNELS) ?
"5.2" : "2.4");
}
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
}
static void
iwm_init_channel_map(struct ieee80211com *ic, int maxchans, int *nchans,
struct ieee80211_channel chans[])
{
struct iwm_softc *sc = ic->ic_softc;
struct iwm_nvm_data *data = &sc->sc_nvm;
Follow-up r298818: hide size of 'bands' array behind a macro.
2016-05-16 02:27:28 +00:00
uint8_t bands[IEEE80211_MODE_BYTES];
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
size_t ch_num;
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
memset(bands, 0, sizeof(bands));
/* 1-13: 11b/g channels. */
setbit(bands, IEEE80211_MODE_11B);
setbit(bands, IEEE80211_MODE_11G);
iwm_add_channel_band(sc, chans, maxchans, nchans, 0,
IWM_NUM_2GHZ_CHANNELS - 1, bands);
/* 14: 11b channel only. */
clrbit(bands, IEEE80211_MODE_11G);
iwm_add_channel_band(sc, chans, maxchans, nchans,
iwm: fix parameters for iwm_add_channel_band() It accepts <first index, max index + 1>, not <first index, number of entries>. Reported by: adrian
2016-05-02 19:07:44 +00:00
IWM_NUM_2GHZ_CHANNELS - 1, IWM_NUM_2GHZ_CHANNELS, bands);
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
if (data->sku_cap_band_52GHz_enable) {
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (sc->sc_device_family == IWM_DEVICE_FAMILY_7000)
ch_num = nitems(iwm_nvm_channels);
else
ch_num = nitems(iwm_nvm_channels_8000);
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
memset(bands, 0, sizeof(bands));
setbit(bands, IEEE80211_MODE_11A);
iwm_add_channel_band(sc, chans, maxchans, nchans,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
IWM_NUM_2GHZ_CHANNELS, ch_num, bands);
}
}
static void
iwm_set_hw_address_8000(struct iwm_softc *sc, struct iwm_nvm_data *data,
const uint16_t *mac_override, const uint16_t *nvm_hw)
{
const uint8_t *hw_addr;
if (mac_override) {
static const uint8_t reserved_mac[] = {
0x02, 0xcc, 0xaa, 0xff, 0xee, 0x00
};
hw_addr = (const uint8_t *)(mac_override +
IWM_MAC_ADDRESS_OVERRIDE_8000);
/*
* Store the MAC address from MAO section.
* No byte swapping is required in MAO section
*/
IEEE80211_ADDR_COPY(data->hw_addr, hw_addr);
/*
* Force the use of the OTP MAC address in case of reserved MAC
* address in the NVM, or if address is given but invalid.
*/
if (!IEEE80211_ADDR_EQ(reserved_mac, hw_addr) &&
!IEEE80211_ADDR_EQ(ieee80211broadcastaddr, data->hw_addr) &&
iwm_is_valid_ether_addr(data->hw_addr) &&
!IEEE80211_IS_MULTICAST(data->hw_addr))
return;
IWM_DPRINTF(sc, IWM_DEBUG_RESET,
"%s: mac address from nvm override section invalid\n",
__func__);
}
if (nvm_hw) {
/* read the mac address from WFMP registers */
uint32_t mac_addr0 =
htole32(iwm_read_prph(sc, IWM_WFMP_MAC_ADDR_0));
uint32_t mac_addr1 =
htole32(iwm_read_prph(sc, IWM_WFMP_MAC_ADDR_1));
hw_addr = (const uint8_t *)&mac_addr0;
data->hw_addr[0] = hw_addr[3];
data->hw_addr[1] = hw_addr[2];
data->hw_addr[2] = hw_addr[1];
data->hw_addr[3] = hw_addr[0];
hw_addr = (const uint8_t *)&mac_addr1;
data->hw_addr[4] = hw_addr[1];
data->hw_addr[5] = hw_addr[0];
return;
}
device_printf(sc->sc_dev, "%s: mac address not found\n", __func__);
memset(data->hw_addr, 0, sizeof(data->hw_addr));
}
static int
iwm_get_sku(const struct iwm_softc *sc, const uint16_t *nvm_sw,
const uint16_t *phy_sku)
{
if (sc->sc_device_family != IWM_DEVICE_FAMILY_8000)
return le16_to_cpup(nvm_sw + IWM_SKU);
return le32_to_cpup((const uint32_t *)(phy_sku + IWM_SKU_8000));
}
static int
iwm_get_nvm_version(const struct iwm_softc *sc, const uint16_t *nvm_sw)
{
if (sc->sc_device_family != IWM_DEVICE_FAMILY_8000)
return le16_to_cpup(nvm_sw + IWM_NVM_VERSION);
else
return le32_to_cpup((const uint32_t *)(nvm_sw +
IWM_NVM_VERSION_8000));
}
static int
iwm_get_radio_cfg(const struct iwm_softc *sc, const uint16_t *nvm_sw,
const uint16_t *phy_sku)
{
if (sc->sc_device_family != IWM_DEVICE_FAMILY_8000)
return le16_to_cpup(nvm_sw + IWM_RADIO_CFG);
return le32_to_cpup((const uint32_t *)(phy_sku + IWM_RADIO_CFG_8000));
}
static int
iwm_get_n_hw_addrs(const struct iwm_softc *sc, const uint16_t *nvm_sw)
{
int n_hw_addr;
if (sc->sc_device_family != IWM_DEVICE_FAMILY_8000)
return le16_to_cpup(nvm_sw + IWM_N_HW_ADDRS);
n_hw_addr = le32_to_cpup((const uint32_t *)(nvm_sw + IWM_N_HW_ADDRS_8000));
return n_hw_addr & IWM_N_HW_ADDR_MASK;
}
static void
iwm_set_radio_cfg(const struct iwm_softc *sc, struct iwm_nvm_data *data,
uint32_t radio_cfg)
{
if (sc->sc_device_family != IWM_DEVICE_FAMILY_8000) {
data->radio_cfg_type = IWM_NVM_RF_CFG_TYPE_MSK(radio_cfg);
data->radio_cfg_step = IWM_NVM_RF_CFG_STEP_MSK(radio_cfg);
data->radio_cfg_dash = IWM_NVM_RF_CFG_DASH_MSK(radio_cfg);
data->radio_cfg_pnum = IWM_NVM_RF_CFG_PNUM_MSK(radio_cfg);
return;
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/* set the radio configuration for family 8000 */
data->radio_cfg_type = IWM_NVM_RF_CFG_TYPE_MSK_8000(radio_cfg);
data->radio_cfg_step = IWM_NVM_RF_CFG_STEP_MSK_8000(radio_cfg);
data->radio_cfg_dash = IWM_NVM_RF_CFG_DASH_MSK_8000(radio_cfg);
data->radio_cfg_pnum = IWM_NVM_RF_CFG_FLAVOR_MSK_8000(radio_cfg);
data->valid_tx_ant = IWM_NVM_RF_CFG_TX_ANT_MSK_8000(radio_cfg);
data->valid_rx_ant = IWM_NVM_RF_CFG_RX_ANT_MSK_8000(radio_cfg);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
static int
iwm_parse_nvm_data(struct iwm_softc *sc,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
const uint16_t *nvm_hw, const uint16_t *nvm_sw,
const uint16_t *nvm_calib, const uint16_t *mac_override,
const uint16_t *phy_sku, const uint16_t *regulatory)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
struct iwm_nvm_data *data = &sc->sc_nvm;
uint8_t hw_addr[IEEE80211_ADDR_LEN];
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
uint32_t sku, radio_cfg;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
data->nvm_version = iwm_get_nvm_version(sc, nvm_sw);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
radio_cfg = iwm_get_radio_cfg(sc, nvm_sw, phy_sku);
iwm_set_radio_cfg(sc, data, radio_cfg);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sku = iwm_get_sku(sc, nvm_sw, phy_sku);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
data->sku_cap_band_24GHz_enable = sku & IWM_NVM_SKU_CAP_BAND_24GHZ;
data->sku_cap_band_52GHz_enable = sku & IWM_NVM_SKU_CAP_BAND_52GHZ;
data->sku_cap_11n_enable = 0;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
data->n_hw_addrs = iwm_get_n_hw_addrs(sc, nvm_sw);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* The byte order is little endian 16 bit, meaning 214365 */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (sc->sc_device_family == IWM_DEVICE_FAMILY_7000) {
IEEE80211_ADDR_COPY(hw_addr, nvm_hw + IWM_HW_ADDR);
data->hw_addr[0] = hw_addr[1];
data->hw_addr[1] = hw_addr[0];
data->hw_addr[2] = hw_addr[3];
data->hw_addr[3] = hw_addr[2];
data->hw_addr[4] = hw_addr[5];
data->hw_addr[5] = hw_addr[4];
} else {
iwm_set_hw_address_8000(sc, data, mac_override, nvm_hw);
}
if (sc->sc_device_family == IWM_DEVICE_FAMILY_7000) {
memcpy(data->nvm_ch_flags, &nvm_sw[IWM_NVM_CHANNELS],
IWM_NUM_CHANNELS * sizeof(uint16_t));
} else {
memcpy(data->nvm_ch_flags, &regulatory[IWM_NVM_CHANNELS_8000],
IWM_NUM_CHANNELS_8000 * sizeof(uint16_t));
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return 0;
}
/*
* END NVM PARSE
*/
static int
iwm_parse_nvm_sections(struct iwm_softc *sc, struct iwm_nvm_section *sections)
{
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
const uint16_t *hw, *sw, *calib, *regulatory, *mac_override, *phy_sku;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Checking for required sections */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (sc->sc_device_family == IWM_DEVICE_FAMILY_7000) {
if (!sections[IWM_NVM_SECTION_TYPE_SW].data ||
!sections[IWM_NVM_SECTION_TYPE_HW].data) {
device_printf(sc->sc_dev,
"Can't parse empty OTP/NVM sections\n");
return ENOENT;
}
hw = (const uint16_t *) sections[IWM_NVM_SECTION_TYPE_HW].data;
} else if (sc->sc_device_family == IWM_DEVICE_FAMILY_8000) {
/* SW and REGULATORY sections are mandatory */
if (!sections[IWM_NVM_SECTION_TYPE_SW].data ||
!sections[IWM_NVM_SECTION_TYPE_REGULATORY].data) {
device_printf(sc->sc_dev,
"Can't parse empty OTP/NVM sections\n");
return ENOENT;
}
/* MAC_OVERRIDE or at least HW section must exist */
if (!sections[IWM_NVM_SECTION_TYPE_HW_8000].data &&
!sections[IWM_NVM_SECTION_TYPE_MAC_OVERRIDE].data) {
device_printf(sc->sc_dev,
"Can't parse mac_address, empty sections\n");
return ENOENT;
}
/* PHY_SKU section is mandatory in B0 */
if (!sections[IWM_NVM_SECTION_TYPE_PHY_SKU].data) {
device_printf(sc->sc_dev,
"Can't parse phy_sku in B0, empty sections\n");
return ENOENT;
}
hw = (const uint16_t *)
sections[IWM_NVM_SECTION_TYPE_HW_8000].data;
} else {
panic("unknown device family %d\n", sc->sc_device_family);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
sw = (const uint16_t *)sections[IWM_NVM_SECTION_TYPE_SW].data;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
calib = (const uint16_t *)
sections[IWM_NVM_SECTION_TYPE_CALIBRATION].data;
regulatory = (const uint16_t *)
sections[IWM_NVM_SECTION_TYPE_REGULATORY].data;
mac_override = (const uint16_t *)
sections[IWM_NVM_SECTION_TYPE_MAC_OVERRIDE].data;
phy_sku = (const uint16_t *)sections[IWM_NVM_SECTION_TYPE_PHY_SKU].data;
return iwm_parse_nvm_data(sc, hw, sw, calib, mac_override,
phy_sku, regulatory);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
static int
iwm_nvm_init(struct iwm_softc *sc)
{
struct iwm_nvm_section nvm_sections[IWM_NVM_NUM_OF_SECTIONS];
int i, section, error;
uint16_t len;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
uint8_t *buf;
const size_t bufsz = IWM_MAX_NVM_SECTION_SIZE;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
memset(nvm_sections, 0 , sizeof(nvm_sections));
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
buf = malloc(bufsz, M_DEVBUF, M_NOWAIT);
if (buf == NULL)
return ENOMEM;
[iwm] free RX ring / NVM memory after they're used. * Free RX ring during detach * Free NVM memory after parsing Tested: * 7260, STA mode Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Obtained from: dragonflybsd Differential Revision: https://reviews.freebsd.org/D6817
2016-06-16 17:59:15 +00:00
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
for (i = 0; i < nitems(nvm_to_read); i++) {
section = nvm_to_read[i];
KASSERT(section <= nitems(nvm_sections),
("too many sections"));
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
error = iwm_nvm_read_section(sc, section, buf, &len, bufsz);
if (error) {
error = 0;
continue;
}
nvm_sections[section].data = malloc(len, M_DEVBUF, M_NOWAIT);
if (nvm_sections[section].data == NULL) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
error = ENOMEM;
break;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
memcpy(nvm_sections[section].data, buf, len);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
nvm_sections[section].length = len;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
free(buf, M_DEVBUF);
[iwm] free RX ring / NVM memory after they're used. * Free RX ring during detach * Free NVM memory after parsing Tested: * 7260, STA mode Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Obtained from: dragonflybsd Differential Revision: https://reviews.freebsd.org/D6817
2016-06-16 17:59:15 +00:00
if (error == 0)
error = iwm_parse_nvm_sections(sc, nvm_sections);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
[iwm] free RX ring / NVM memory after they're used. * Free RX ring during detach * Free NVM memory after parsing Tested: * 7260, STA mode Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Obtained from: dragonflybsd Differential Revision: https://reviews.freebsd.org/D6817
2016-06-16 17:59:15 +00:00
for (i = 0; i < IWM_NVM_NUM_OF_SECTIONS; i++) {
if (nvm_sections[i].data != NULL)
free(nvm_sections[i].data, M_DEVBUF);
}
return error;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
/*
* Firmware loading gunk. This is kind of a weird hybrid between the
* iwn driver and the Linux iwlwifi driver.
*/
static int
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
iwm_firmware_load_sect(struct iwm_softc *sc, uint32_t dst_addr,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
const uint8_t *section, uint32_t byte_cnt)
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
{
int error = EINVAL;
uint32_t chunk_sz, offset;
chunk_sz = MIN(IWM_FH_MEM_TB_MAX_LENGTH, byte_cnt);
for (offset = 0; offset < byte_cnt; offset += chunk_sz) {
uint32_t addr, len;
const uint8_t *data;
addr = dst_addr + offset;
len = MIN(chunk_sz, byte_cnt - offset);
data = section + offset;
error = iwm_firmware_load_chunk(sc, addr, data, len);
if (error)
break;
}
return error;
}
static int
iwm_firmware_load_chunk(struct iwm_softc *sc, uint32_t dst_addr,
const uint8_t *chunk, uint32_t byte_cnt)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
struct iwm_dma_info *dma = &sc->fw_dma;
int error;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/* Copy firmware chunk into pre-allocated DMA-safe memory. */
memcpy(dma->vaddr, chunk, byte_cnt);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
bus_dmamap_sync(dma->tag, dma->map, BUS_DMASYNC_PREWRITE);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (dst_addr >= IWM_FW_MEM_EXTENDED_START &&
dst_addr <= IWM_FW_MEM_EXTENDED_END) {
iwm_set_bits_prph(sc, IWM_LMPM_CHICK,
IWM_LMPM_CHICK_EXTENDED_ADDR_SPACE);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
sc->sc_fw_chunk_done = 0;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (!iwm_nic_lock(sc))
return EBUSY;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_WRITE(sc, IWM_FH_TCSR_CHNL_TX_CONFIG_REG(IWM_FH_SRVC_CHNL),
IWM_FH_TCSR_TX_CONFIG_REG_VAL_DMA_CHNL_PAUSE);
IWM_WRITE(sc, IWM_FH_SRVC_CHNL_SRAM_ADDR_REG(IWM_FH_SRVC_CHNL),
dst_addr);
IWM_WRITE(sc, IWM_FH_TFDIB_CTRL0_REG(IWM_FH_SRVC_CHNL),
dma->paddr & IWM_FH_MEM_TFDIB_DRAM_ADDR_LSB_MSK);
IWM_WRITE(sc, IWM_FH_TFDIB_CTRL1_REG(IWM_FH_SRVC_CHNL),
(iwm_get_dma_hi_addr(dma->paddr)
<< IWM_FH_MEM_TFDIB_REG1_ADDR_BITSHIFT) | byte_cnt);
IWM_WRITE(sc, IWM_FH_TCSR_CHNL_TX_BUF_STS_REG(IWM_FH_SRVC_CHNL),
1 << IWM_FH_TCSR_CHNL_TX_BUF_STS_REG_POS_TB_NUM |
1 << IWM_FH_TCSR_CHNL_TX_BUF_STS_REG_POS_TB_IDX |
IWM_FH_TCSR_CHNL_TX_BUF_STS_REG_VAL_TFDB_VALID);
IWM_WRITE(sc, IWM_FH_TCSR_CHNL_TX_CONFIG_REG(IWM_FH_SRVC_CHNL),
IWM_FH_TCSR_TX_CONFIG_REG_VAL_DMA_CHNL_ENABLE |
IWM_FH_TCSR_TX_CONFIG_REG_VAL_DMA_CREDIT_DISABLE |
IWM_FH_TCSR_TX_CONFIG_REG_VAL_CIRQ_HOST_ENDTFD);
iwm_nic_unlock(sc);
/* wait 1s for this segment to load */
while (!sc->sc_fw_chunk_done)
if ((error = msleep(&sc->sc_fw, &sc->sc_mtx, 0, "iwmfw", hz)) != 0)
break;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (!sc->sc_fw_chunk_done) {
device_printf(sc->sc_dev,
"fw chunk addr 0x%x len %d failed to load\n",
dst_addr, byte_cnt);
}
if (dst_addr >= IWM_FW_MEM_EXTENDED_START &&
dst_addr <= IWM_FW_MEM_EXTENDED_END && iwm_nic_lock(sc)) {
iwm_clear_bits_prph(sc, IWM_LMPM_CHICK,
IWM_LMPM_CHICK_EXTENDED_ADDR_SPACE);
iwm_nic_unlock(sc);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
int
iwm_load_cpu_sections_8000(struct iwm_softc *sc, struct iwm_fw_sects *fws,
int cpu, int *first_ucode_section)
{
int shift_param;
int i, error = 0, sec_num = 0x1;
uint32_t val, last_read_idx = 0;
const void *data;
uint32_t dlen;
uint32_t offset;
if (cpu == 1) {
shift_param = 0;
*first_ucode_section = 0;
} else {
shift_param = 16;
(*first_ucode_section)++;
}
for (i = *first_ucode_section; i < IWM_UCODE_SECT_MAX; i++) {
last_read_idx = i;
data = fws->fw_sect[i].fws_data;
dlen = fws->fw_sect[i].fws_len;
offset = fws->fw_sect[i].fws_devoff;
/*
* CPU1_CPU2_SEPARATOR_SECTION delimiter - separate between
* CPU1 to CPU2.
* PAGING_SEPARATOR_SECTION delimiter - separate between
* CPU2 non paged to CPU2 paging sec.
*/
if (!data || offset == IWM_CPU1_CPU2_SEPARATOR_SECTION ||
offset == IWM_PAGING_SEPARATOR_SECTION)
break;
IWM_DPRINTF(sc, IWM_DEBUG_RESET,
"LOAD FIRMWARE chunk %d offset 0x%x len %d for cpu %d\n",
i, offset, dlen, cpu);
if (dlen > sc->sc_fwdmasegsz) {
IWM_DPRINTF(sc, IWM_DEBUG_RESET,
"chunk %d too large (%d bytes)\n", i, dlen);
error = EFBIG;
} else {
error = iwm_firmware_load_sect(sc, offset, data, dlen);
}
if (error) {
device_printf(sc->sc_dev,
"could not load firmware chunk %d (error %d)\n",
i, error);
return error;
}
/* Notify the ucode of the loaded section number and status */
if (iwm_nic_lock(sc)) {
val = IWM_READ(sc, IWM_FH_UCODE_LOAD_STATUS);
val = val | (sec_num << shift_param);
IWM_WRITE(sc, IWM_FH_UCODE_LOAD_STATUS, val);
sec_num = (sec_num << 1) | 0x1;
iwm_nic_unlock(sc);
/*
* The firmware won't load correctly without this delay.
*/
DELAY(8000);
}
}
*first_ucode_section = last_read_idx;
if (iwm_nic_lock(sc)) {
if (cpu == 1)
IWM_WRITE(sc, IWM_FH_UCODE_LOAD_STATUS, 0xFFFF);
else
IWM_WRITE(sc, IWM_FH_UCODE_LOAD_STATUS, 0xFFFFFFFF);
iwm_nic_unlock(sc);
}
return 0;
}
int
iwm_load_firmware_8000(struct iwm_softc *sc, enum iwm_ucode_type ucode_type)
{
struct iwm_fw_sects *fws;
int error = 0;
int first_ucode_section;
IWM_DPRINTF(sc, IWM_DEBUG_RESET, "loading ucode type %d\n",
ucode_type);
fws = &sc->sc_fw.fw_sects[ucode_type];
/* configure the ucode to be ready to get the secured image */
/* release CPU reset */
iwm_write_prph(sc, IWM_RELEASE_CPU_RESET, IWM_RELEASE_CPU_RESET_BIT);
/* load to FW the binary Secured sections of CPU1 */
error = iwm_load_cpu_sections_8000(sc, fws, 1, &first_ucode_section);
if (error)
return error;
/* load to FW the binary sections of CPU2 */
return iwm_load_cpu_sections_8000(sc, fws, 2, &first_ucode_section);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
iwm_load_firmware_7000(struct iwm_softc *sc, enum iwm_ucode_type ucode_type)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
struct iwm_fw_sects *fws;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
int error, i;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
const void *data;
uint32_t dlen;
uint32_t offset;
sc->sc_uc.uc_intr = 0;
fws = &sc->sc_fw.fw_sects[ucode_type];
for (i = 0; i < fws->fw_count; i++) {
data = fws->fw_sect[i].fws_data;
dlen = fws->fw_sect[i].fws_len;
offset = fws->fw_sect[i].fws_devoff;
IWM_DPRINTF(sc, IWM_DEBUG_FIRMWARE_TLV,
"LOAD FIRMWARE type %d offset %u len %d\n",
ucode_type, offset, dlen);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (dlen > sc->sc_fwdmasegsz) {
IWM_DPRINTF(sc, IWM_DEBUG_FIRMWARE_TLV,
"chunk %d too large (%d bytes)\n", i, dlen);
error = EFBIG;
} else {
error = iwm_firmware_load_sect(sc, offset, data, dlen);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (error) {
device_printf(sc->sc_dev,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
"could not load firmware chunk %u of %u "
"(error=%d)\n", i, fws->fw_count, error);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
}
}
IWM_WRITE(sc, IWM_CSR_RESET, 0);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
return 0;
}
static int
iwm_load_firmware(struct iwm_softc *sc, enum iwm_ucode_type ucode_type)
{
int error, w;
if (sc->sc_device_family == IWM_DEVICE_FAMILY_8000)
error = iwm_load_firmware_8000(sc, ucode_type);
else
error = iwm_load_firmware_7000(sc, ucode_type);
if (error)
return error;
/* wait for the firmware to load */
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
for (w = 0; !sc->sc_uc.uc_intr && w < 10; w++) {
error = msleep(&sc->sc_uc, &sc->sc_mtx, 0, "iwmuc", hz/10);
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (error || !sc->sc_uc.uc_ok) {
device_printf(sc->sc_dev, "could not load firmware\n");
if (sc->sc_device_family == IWM_DEVICE_FAMILY_8000) {
device_printf(sc->sc_dev, "cpu1 status: 0x%x\n",
iwm_read_prph(sc, IWM_SB_CPU_1_STATUS));
device_printf(sc->sc_dev, "cpu2 status: 0x%x\n",
iwm_read_prph(sc, IWM_SB_CPU_2_STATUS));
}
}
/*
* Give the firmware some time to initialize.
* Accessing it too early causes errors.
*/
msleep(&w, &sc->sc_mtx, 0, "iwmfwinit", hz);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
}
/* iwlwifi: pcie/trans.c */
static int
iwm_start_fw(struct iwm_softc *sc, enum iwm_ucode_type ucode_type)
{
int error;
IWM_WRITE(sc, IWM_CSR_INT, ~0);
if ((error = iwm_nic_init(sc)) != 0) {
device_printf(sc->sc_dev, "unable to init nic\n");
return error;
}
/* make sure rfkill handshake bits are cleared */
IWM_WRITE(sc, IWM_CSR_UCODE_DRV_GP1_CLR, IWM_CSR_UCODE_SW_BIT_RFKILL);
IWM_WRITE(sc, IWM_CSR_UCODE_DRV_GP1_CLR,
IWM_CSR_UCODE_DRV_GP1_BIT_CMD_BLOCKED);
/* clear (again), then enable host interrupts */
IWM_WRITE(sc, IWM_CSR_INT, ~0);
iwm_enable_interrupts(sc);
/* really make sure rfkill handshake bits are cleared */
/* maybe we should write a few times more? just to make sure */
IWM_WRITE(sc, IWM_CSR_UCODE_DRV_GP1_CLR, IWM_CSR_UCODE_SW_BIT_RFKILL);
IWM_WRITE(sc, IWM_CSR_UCODE_DRV_GP1_CLR, IWM_CSR_UCODE_SW_BIT_RFKILL);
/* Load the given image to the HW */
return iwm_load_firmware(sc, ucode_type);
}
static int
iwm_send_tx_ant_cfg(struct iwm_softc *sc, uint8_t valid_tx_ant)
{
struct iwm_tx_ant_cfg_cmd tx_ant_cmd = {
.valid = htole32(valid_tx_ant),
};
return iwm_mvm_send_cmd_pdu(sc, IWM_TX_ANT_CONFIGURATION_CMD,
IWM_CMD_SYNC, sizeof(tx_ant_cmd), &tx_ant_cmd);
}
/* iwlwifi: mvm/fw.c */
static int
iwm_send_phy_cfg_cmd(struct iwm_softc *sc)
{
struct iwm_phy_cfg_cmd phy_cfg_cmd;
enum iwm_ucode_type ucode_type = sc->sc_uc_current;
/* Set parameters */
phy_cfg_cmd.phy_cfg = htole32(sc->sc_fw_phy_config);
phy_cfg_cmd.calib_control.event_trigger =
sc->sc_default_calib[ucode_type].event_trigger;
phy_cfg_cmd.calib_control.flow_trigger =
sc->sc_default_calib[ucode_type].flow_trigger;
IWM_DPRINTF(sc, IWM_DEBUG_CMD | IWM_DEBUG_RESET,
"Sending Phy CFG command: 0x%x\n", phy_cfg_cmd.phy_cfg);
return iwm_mvm_send_cmd_pdu(sc, IWM_PHY_CONFIGURATION_CMD, IWM_CMD_SYNC,
sizeof(phy_cfg_cmd), &phy_cfg_cmd);
}
static int
iwm_mvm_load_ucode_wait_alive(struct iwm_softc *sc,
enum iwm_ucode_type ucode_type)
{
enum iwm_ucode_type old_type = sc->sc_uc_current;
int error;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if ((error = iwm_read_firmware(sc, ucode_type)) != 0) {
device_printf(sc->sc_dev, "iwm_read_firmware: failed %d\n",
error);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
sc->sc_uc_current = ucode_type;
error = iwm_start_fw(sc, ucode_type);
if (error) {
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
device_printf(sc->sc_dev, "iwm_start_fw: failed %d\n", error);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
sc->sc_uc_current = old_type;
return error;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
error = iwm_post_alive(sc);
if (error) {
device_printf(sc->sc_dev, "iwm_fw_alive: failed %d\n", error);
}
return error;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
/*
* mvm misc bits
*/
/*
* follows iwlwifi/fw.c
*/
static int
iwm_run_init_mvm_ucode(struct iwm_softc *sc, int justnvm)
{
int error;
/* do not operate with rfkill switch turned on */
if ((sc->sc_flags & IWM_FLAG_RFKILL) && !justnvm) {
device_printf(sc->sc_dev,
"radio is disabled by hardware switch\n");
return EPERM;
}
sc->sc_init_complete = 0;
if ((error = iwm_mvm_load_ucode_wait_alive(sc,
[iwm] Fix several nitpicks in iwm(4). Move some declarations to if_iwmreg.h. Remove iwm_fw_alive(); just call iwm_post_alive() directly. Simplify iwm_mvm_add_sta(). Return timeout error from iwm_apm_init(). Print a message when init (i.e. boot) firmware fails to load. Remove some commented-out code which wouldn't compile anyway. Move iwm_mvm_tx_fifo to if_iwmreg.h to match better where Linux puts it. Taken-From: OpenBSD (if_iwm.c r1.80 and if_iwmreg.h r1.11) Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD 29fcb331e5620ae145a6ab9cdda830e22fff626a
2016-06-02 05:43:16 +00:00
IWM_UCODE_TYPE_INIT)) != 0) {
device_printf(sc->sc_dev, "failed to load init firmware\n");
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
[iwm] Fix several nitpicks in iwm(4). Move some declarations to if_iwmreg.h. Remove iwm_fw_alive(); just call iwm_post_alive() directly. Simplify iwm_mvm_add_sta(). Return timeout error from iwm_apm_init(). Print a message when init (i.e. boot) firmware fails to load. Remove some commented-out code which wouldn't compile anyway. Move iwm_mvm_tx_fifo to if_iwmreg.h to match better where Linux puts it. Taken-From: OpenBSD (if_iwm.c r1.80 and if_iwmreg.h r1.11) Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD 29fcb331e5620ae145a6ab9cdda830e22fff626a
2016-06-02 05:43:16 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (justnvm) {
if ((error = iwm_nvm_init(sc)) != 0) {
device_printf(sc->sc_dev, "failed to read nvm\n");
return error;
}
Fix IEEE80211_ADDR_COPY() usage. Reviewed by: adrian
2015-10-22 01:36:16 +00:00
IEEE80211_ADDR_COPY(sc->sc_ic.ic_macaddr, sc->sc_nvm.hw_addr);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return 0;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if ((error = iwm_send_bt_init_conf(sc)) != 0) {
device_printf(sc->sc_dev,
"failed to send bt coex configuration: %d\n", error);
return error;
}
/* Init Smart FIFO. */
error = iwm_mvm_sf_config(sc, IWM_SF_INIT_OFF);
if (error != 0)
return error;
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_RESET,
"%s: phy_txant=0x%08x, nvm_valid_tx_ant=0x%02x, valid=0x%02x\n",
__func__,
((sc->sc_fw_phy_config & IWM_FW_PHY_CFG_TX_CHAIN)
>> IWM_FW_PHY_CFG_TX_CHAIN_POS),
sc->sc_nvm.valid_tx_ant,
iwm_fw_valid_tx_ant(sc));
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Send TX valid antennas before triggering calibrations */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if ((error = iwm_send_tx_ant_cfg(sc, iwm_fw_valid_tx_ant(sc))) != 0) {
device_printf(sc->sc_dev,
"failed to send antennas before calibration: %d\n", error);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
* Send phy configurations command to init uCode
* to start the 16.0 uCode init image internal calibrations.
*/
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if ((error = iwm_send_phy_cfg_cmd(sc)) != 0 ) {
device_printf(sc->sc_dev,
"%s: failed to run internal calibration: %d\n",
__func__, error);
return error;
}
/*
* Nothing to do but wait for the init complete notification
* from the firmware
*/
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
while (!sc->sc_init_complete) {
error = msleep(&sc->sc_init_complete, &sc->sc_mtx,
0, "iwminit", 2*hz);
if (error) {
device_printf(sc->sc_dev, "init complete failed: %d\n",
sc->sc_init_complete);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
break;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
}
IWM_DPRINTF(sc, IWM_DEBUG_RESET, "init %scomplete\n",
sc->sc_init_complete ? "" : "not ");
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
}
/*
* receive side
*/
/* (re)stock rx ring, called at init-time and at runtime */
static int
iwm_rx_addbuf(struct iwm_softc *sc, int size, int idx)
{
struct iwm_rx_ring *ring = &sc->rxq;
struct iwm_rx_data *data = &ring->data[idx];
struct mbuf *m;
[iwm] Fix up busdma use in the RX path When allocating a new mbuf or bus_dmamap_load()-ing it fails, we can just keep the old mbuf since we are dropping that packet anyway. Instead of doing bus_dmamap_create() and bus_dmamap_destroy() all the time, create an extra bus_dmamap_t which we can use to safely try bus_dmamap_load()-ing the new mbuf. On success we just swap the spare bus_dmamap_t with the data->map of that ring entry. Tested: Tested with Intel AC7260, verified with vmstat -m that new kernel no longer visibly leaks memory from the M_DEVBUF malloc type. Before, leakage was 1KB every few seconds while ping(8)-ing over the wlan connection. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re@ Obtained from: DragonflyBSD.git cc440b26818b5dfdd9af504d71c1b0e6522b53ef Differential Revision: https://reviews.freebsd.org/D6742
2016-06-13 00:13:20 +00:00
bus_dmamap_t dmamap = NULL;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
bus_dma_segment_t seg;
int nsegs, error;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
m = m_getjcl(M_NOWAIT, MT_DATA, M_PKTHDR, IWM_RBUF_SIZE);
if (m == NULL)
return ENOBUFS;
m->m_len = m->m_pkthdr.len = m->m_ext.ext_size;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
error = bus_dmamap_load_mbuf_sg(ring->data_dmat, ring->spare_map, m,
&seg, &nsegs, BUS_DMA_NOWAIT);
if (error != 0) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev,
[iwm] Fix up busdma use in the RX path When allocating a new mbuf or bus_dmamap_load()-ing it fails, we can just keep the old mbuf since we are dropping that packet anyway. Instead of doing bus_dmamap_create() and bus_dmamap_destroy() all the time, create an extra bus_dmamap_t which we can use to safely try bus_dmamap_load()-ing the new mbuf. On success we just swap the spare bus_dmamap_t with the data->map of that ring entry. Tested: Tested with Intel AC7260, verified with vmstat -m that new kernel no longer visibly leaks memory from the M_DEVBUF malloc type. Before, leakage was 1KB every few seconds while ping(8)-ing over the wlan connection. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re@ Obtained from: DragonflyBSD.git cc440b26818b5dfdd9af504d71c1b0e6522b53ef Differential Revision: https://reviews.freebsd.org/D6742
2016-06-13 00:13:20 +00:00
"%s: can't map mbuf, error %d\n", __func__, error);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
goto fail;
}
[iwm] Fix up busdma use in the RX path When allocating a new mbuf or bus_dmamap_load()-ing it fails, we can just keep the old mbuf since we are dropping that packet anyway. Instead of doing bus_dmamap_create() and bus_dmamap_destroy() all the time, create an extra bus_dmamap_t which we can use to safely try bus_dmamap_load()-ing the new mbuf. On success we just swap the spare bus_dmamap_t with the data->map of that ring entry. Tested: Tested with Intel AC7260, verified with vmstat -m that new kernel no longer visibly leaks memory from the M_DEVBUF malloc type. Before, leakage was 1KB every few seconds while ping(8)-ing over the wlan connection. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re@ Obtained from: DragonflyBSD.git cc440b26818b5dfdd9af504d71c1b0e6522b53ef Differential Revision: https://reviews.freebsd.org/D6742
2016-06-13 00:13:20 +00:00
if (data->m != NULL)
bus_dmamap_unload(ring->data_dmat, data->map);
/* Swap ring->spare_map with data->map */
dmamap = data->map;
data->map = ring->spare_map;
ring->spare_map = dmamap;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
bus_dmamap_sync(ring->data_dmat, data->map, BUS_DMASYNC_PREREAD);
[iwm] Fix up busdma use in the RX path When allocating a new mbuf or bus_dmamap_load()-ing it fails, we can just keep the old mbuf since we are dropping that packet anyway. Instead of doing bus_dmamap_create() and bus_dmamap_destroy() all the time, create an extra bus_dmamap_t which we can use to safely try bus_dmamap_load()-ing the new mbuf. On success we just swap the spare bus_dmamap_t with the data->map of that ring entry. Tested: Tested with Intel AC7260, verified with vmstat -m that new kernel no longer visibly leaks memory from the M_DEVBUF malloc type. Before, leakage was 1KB every few seconds while ping(8)-ing over the wlan connection. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re@ Obtained from: DragonflyBSD.git cc440b26818b5dfdd9af504d71c1b0e6522b53ef Differential Revision: https://reviews.freebsd.org/D6742
2016-06-13 00:13:20 +00:00
data->m = m;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Update RX descriptor. */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
KASSERT((seg.ds_addr & 255) == 0, ("seg.ds_addr not aligned"));
ring->desc[idx] = htole32(seg.ds_addr >> 8);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
bus_dmamap_sync(ring->desc_dma.tag, ring->desc_dma.map,
BUS_DMASYNC_PREWRITE);
return 0;
fail:
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
m_freem(m);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
}
/* iwlwifi: mvm/rx.c */
#define IWM_RSSI_OFFSET 50
static int
iwm_mvm_calc_rssi(struct iwm_softc *sc, struct iwm_rx_phy_info *phy_info)
{
int rssi_a, rssi_b, rssi_a_dbm, rssi_b_dbm, max_rssi_dbm;
uint32_t agc_a, agc_b;
uint32_t val;
val = le32toh(phy_info->non_cfg_phy[IWM_RX_INFO_AGC_IDX]);
agc_a = (val & IWM_OFDM_AGC_A_MSK) >> IWM_OFDM_AGC_A_POS;
agc_b = (val & IWM_OFDM_AGC_B_MSK) >> IWM_OFDM_AGC_B_POS;
val = le32toh(phy_info->non_cfg_phy[IWM_RX_INFO_RSSI_AB_IDX]);
rssi_a = (val & IWM_OFDM_RSSI_INBAND_A_MSK) >> IWM_OFDM_RSSI_A_POS;
rssi_b = (val & IWM_OFDM_RSSI_INBAND_B_MSK) >> IWM_OFDM_RSSI_B_POS;
/*
* dBm = rssi dB - agc dB - constant.
* Higher AGC (higher radio gain) means lower signal.
*/
rssi_a_dbm = rssi_a - IWM_RSSI_OFFSET - agc_a;
rssi_b_dbm = rssi_b - IWM_RSSI_OFFSET - agc_b;
max_rssi_dbm = MAX(rssi_a_dbm, rssi_b_dbm);
IWM_DPRINTF(sc, IWM_DEBUG_RECV,
"Rssi In A %d B %d Max %d AGCA %d AGCB %d\n",
rssi_a_dbm, rssi_b_dbm, max_rssi_dbm, agc_a, agc_b);
return max_rssi_dbm;
}
/* iwlwifi: mvm/rx.c */
/*
* iwm_mvm_get_signal_strength - use new rx PHY INFO API
* values are reported by the fw as positive values - need to negate
* to obtain their dBM. Account for missing antennas by replacing 0
* values by -256dBm: practically 0 power and a non-feasible 8 bit value.
*/
static int
iwm_mvm_get_signal_strength(struct iwm_softc *sc, struct iwm_rx_phy_info *phy_info)
{
int energy_a, energy_b, energy_c, max_energy;
uint32_t val;
val = le32toh(phy_info->non_cfg_phy[IWM_RX_INFO_ENERGY_ANT_ABC_IDX]);
energy_a = (val & IWM_RX_INFO_ENERGY_ANT_A_MSK) >>
IWM_RX_INFO_ENERGY_ANT_A_POS;
energy_a = energy_a ? -energy_a : -256;
energy_b = (val & IWM_RX_INFO_ENERGY_ANT_B_MSK) >>
IWM_RX_INFO_ENERGY_ANT_B_POS;
energy_b = energy_b ? -energy_b : -256;
energy_c = (val & IWM_RX_INFO_ENERGY_ANT_C_MSK) >>
IWM_RX_INFO_ENERGY_ANT_C_POS;
energy_c = energy_c ? -energy_c : -256;
max_energy = MAX(energy_a, energy_b);
max_energy = MAX(max_energy, energy_c);
IWM_DPRINTF(sc, IWM_DEBUG_RECV,
"energy In A %d B %d C %d , and max %d\n",
energy_a, energy_b, energy_c, max_energy);
return max_energy;
}
static void
iwm_mvm_rx_rx_phy_cmd(struct iwm_softc *sc,
struct iwm_rx_packet *pkt, struct iwm_rx_data *data)
{
struct iwm_rx_phy_info *phy_info = (void *)pkt->data;
IWM_DPRINTF(sc, IWM_DEBUG_RECV, "received PHY stats\n");
bus_dmamap_sync(sc->rxq.data_dmat, data->map, BUS_DMASYNC_POSTREAD);
memcpy(&sc->sc_last_phy_info, phy_info, sizeof(sc->sc_last_phy_info));
}
/*
* Retrieve the average noise (in dBm) among receivers.
*/
static int
[iwm] fix up RSSI calculations for both scan results and normal RX operations. * hard code a noise floor of -96 for now. The noise floor update code returns some "interesting" values that I can't map to anything useful right now. * Ensure a default noise floor is set - otherwise the initial scan results have a noise floor of '0'. * Fix up the RSSI calculation to be correctly relative to the noise floor. The RSSI routines return an absolute value in dBm - so fix this up. * Cap RSSI values appropriately. * Ensure we pass in a 1/2 dB unit value in to net80211. Tested: * Intel 7260, STA mode iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: hw rev 0x140, fw ver 16.242414.0, address xx:xx:xx:xx:xx:xx
2016-09-18 05:07:18 +00:00
iwm_get_noise(struct iwm_softc *sc,
const struct iwm_mvm_statistics_rx_non_phy *stats)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
int i, total, nbant, noise;
total = nbant = noise = 0;
for (i = 0; i < 3; i++) {
noise = le32toh(stats->beacon_silence_rssi[i]) & 0xff;
[iwm] fix up RSSI calculations for both scan results and normal RX operations. * hard code a noise floor of -96 for now. The noise floor update code returns some "interesting" values that I can't map to anything useful right now. * Ensure a default noise floor is set - otherwise the initial scan results have a noise floor of '0'. * Fix up the RSSI calculation to be correctly relative to the noise floor. The RSSI routines return an absolute value in dBm - so fix this up. * Cap RSSI values appropriately. * Ensure we pass in a 1/2 dB unit value in to net80211. Tested: * Intel 7260, STA mode iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: hw rev 0x140, fw ver 16.242414.0, address xx:xx:xx:xx:xx:xx
2016-09-18 05:07:18 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_RECV, "%s: i=%d, noise=%d\n",
__func__,
i,
noise);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (noise) {
total += noise;
nbant++;
}
}
[iwm] fix up RSSI calculations for both scan results and normal RX operations. * hard code a noise floor of -96 for now. The noise floor update code returns some "interesting" values that I can't map to anything useful right now. * Ensure a default noise floor is set - otherwise the initial scan results have a noise floor of '0'. * Fix up the RSSI calculation to be correctly relative to the noise floor. The RSSI routines return an absolute value in dBm - so fix this up. * Cap RSSI values appropriately. * Ensure we pass in a 1/2 dB unit value in to net80211. Tested: * Intel 7260, STA mode iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: hw rev 0x140, fw ver 16.242414.0, address xx:xx:xx:xx:xx:xx
2016-09-18 05:07:18 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_RECV, "%s: nbant=%d, total=%d\n",
__func__, nbant, total);
#if 0
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* There should be at least one antenna but check anyway. */
return (nbant == 0) ? -127 : (total / nbant) - 107;
[iwm] fix up RSSI calculations for both scan results and normal RX operations. * hard code a noise floor of -96 for now. The noise floor update code returns some "interesting" values that I can't map to anything useful right now. * Ensure a default noise floor is set - otherwise the initial scan results have a noise floor of '0'. * Fix up the RSSI calculation to be correctly relative to the noise floor. The RSSI routines return an absolute value in dBm - so fix this up. * Cap RSSI values appropriately. * Ensure we pass in a 1/2 dB unit value in to net80211. Tested: * Intel 7260, STA mode iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: hw rev 0x140, fw ver 16.242414.0, address xx:xx:xx:xx:xx:xx
2016-09-18 05:07:18 +00:00
#else
/* For now, just hard-code it to -96 to be safe */
return (-96);
#endif
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
/*
* iwm_mvm_rx_rx_mpdu - IWM_REPLY_RX_MPDU_CMD handler
*
* Handles the actual data of the Rx packet from the fw
*/
static void
iwm_mvm_rx_rx_mpdu(struct iwm_softc *sc,
struct iwm_rx_packet *pkt, struct iwm_rx_data *data)
{
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct ieee80211com *ic = &sc->sc_ic;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
struct ieee80211vap *vap = TAILQ_FIRST(&ic->ic_vaps);
struct ieee80211_frame *wh;
struct ieee80211_node *ni;
struct ieee80211_rx_stats rxs;
struct mbuf *m;
struct iwm_rx_phy_info *phy_info;
struct iwm_rx_mpdu_res_start *rx_res;
uint32_t len;
uint32_t rx_pkt_status;
int rssi;
bus_dmamap_sync(sc->rxq.data_dmat, data->map, BUS_DMASYNC_POSTREAD);
phy_info = &sc->sc_last_phy_info;
rx_res = (struct iwm_rx_mpdu_res_start *)pkt->data;
wh = (struct ieee80211_frame *)(pkt->data + sizeof(*rx_res));
len = le16toh(rx_res->byte_count);
rx_pkt_status = le32toh(*(uint32_t *)(pkt->data + sizeof(*rx_res) + len));
m = data->m;
m->m_data = pkt->data + sizeof(*rx_res);
m->m_pkthdr.len = m->m_len = len;
if (__predict_false(phy_info->cfg_phy_cnt > 20)) {
device_printf(sc->sc_dev,
"dsp size out of range [0,20]: %d\n",
phy_info->cfg_phy_cnt);
iwm: add missing 'Rx errors' counter incrementation.
2016-10-09 10:21:28 +00:00
goto fail;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
if (!(rx_pkt_status & IWM_RX_MPDU_RES_STATUS_CRC_OK) ||
!(rx_pkt_status & IWM_RX_MPDU_RES_STATUS_OVERRUN_OK)) {
IWM_DPRINTF(sc, IWM_DEBUG_RECV,
"Bad CRC or FIFO: 0x%08X.\n", rx_pkt_status);
iwm: add missing 'Rx errors' counter incrementation.
2016-10-09 10:21:28 +00:00
goto fail;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
if (sc->sc_capaflags & IWM_UCODE_TLV_FLAGS_RX_ENERGY_API) {
rssi = iwm_mvm_get_signal_strength(sc, phy_info);
} else {
rssi = iwm_mvm_calc_rssi(sc, phy_info);
}
[iwm] fix up RSSI calculations for both scan results and normal RX operations. * hard code a noise floor of -96 for now. The noise floor update code returns some "interesting" values that I can't map to anything useful right now. * Ensure a default noise floor is set - otherwise the initial scan results have a noise floor of '0'. * Fix up the RSSI calculation to be correctly relative to the noise floor. The RSSI routines return an absolute value in dBm - so fix this up. * Cap RSSI values appropriately. * Ensure we pass in a 1/2 dB unit value in to net80211. Tested: * Intel 7260, STA mode iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: hw rev 0x140, fw ver 16.242414.0, address xx:xx:xx:xx:xx:xx
2016-09-18 05:07:18 +00:00
/* Note: RSSI is absolute (ie a -ve value) */
if (rssi < IWM_MIN_DBM)
rssi = IWM_MIN_DBM;
else if (rssi > IWM_MAX_DBM)
rssi = IWM_MAX_DBM;
/* Map it to relative value */
rssi = rssi - sc->sc_noise;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* replenish ring for the buffer we're going to feed to the sharks */
if (iwm_rx_addbuf(sc, IWM_RBUF_SIZE, sc->rxq.cur) != 0) {
device_printf(sc->sc_dev, "%s: unable to add more buffers\n",
__func__);
iwm: add missing 'Rx errors' counter incrementation.
2016-10-09 10:21:28 +00:00
goto fail;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
[iwm] fix up RSSI calculations for both scan results and normal RX operations. * hard code a noise floor of -96 for now. The noise floor update code returns some "interesting" values that I can't map to anything useful right now. * Ensure a default noise floor is set - otherwise the initial scan results have a noise floor of '0'. * Fix up the RSSI calculation to be correctly relative to the noise floor. The RSSI routines return an absolute value in dBm - so fix this up. * Cap RSSI values appropriately. * Ensure we pass in a 1/2 dB unit value in to net80211. Tested: * Intel 7260, STA mode iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: hw rev 0x140, fw ver 16.242414.0, address xx:xx:xx:xx:xx:xx
2016-09-18 05:07:18 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_RECV,
"%s: rssi=%d, noise=%d\n", __func__, rssi, sc->sc_noise);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ni = ieee80211_find_rxnode(ic, (struct ieee80211_frame_min *)wh);
IWM_DPRINTF(sc, IWM_DEBUG_RECV,
"%s: phy_info: channel=%d, flags=0x%08x\n",
__func__,
le16toh(phy_info->channel),
le16toh(phy_info->phy_flags));
/*
* Populate an RX state struct with the provided information.
*/
bzero(&rxs, sizeof(rxs));
rxs.r_flags |= IEEE80211_R_IEEE | IEEE80211_R_FREQ;
rxs.r_flags |= IEEE80211_R_NF | IEEE80211_R_RSSI;
rxs.c_ieee = le16toh(phy_info->channel);
if (le16toh(phy_info->phy_flags & IWM_RX_RES_PHY_FLAGS_BAND_24)) {
rxs.c_freq = ieee80211_ieee2mhz(rxs.c_ieee, IEEE80211_CHAN_2GHZ);
} else {
rxs.c_freq = ieee80211_ieee2mhz(rxs.c_ieee, IEEE80211_CHAN_5GHZ);
}
[iwm] fix up RSSI calculations for both scan results and normal RX operations. * hard code a noise floor of -96 for now. The noise floor update code returns some "interesting" values that I can't map to anything useful right now. * Ensure a default noise floor is set - otherwise the initial scan results have a noise floor of '0'. * Fix up the RSSI calculation to be correctly relative to the noise floor. The RSSI routines return an absolute value in dBm - so fix this up. * Cap RSSI values appropriately. * Ensure we pass in a 1/2 dB unit value in to net80211. Tested: * Intel 7260, STA mode iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: hw rev 0x140, fw ver 16.242414.0, address xx:xx:xx:xx:xx:xx
2016-09-18 05:07:18 +00:00
/* rssi is in 1/2db units */
[net80211] extend the ieee80211_rx_stats struct to include more information. There are a variety of more interesting RX statistics that we should keep track of but we don't. This is a starting point for adding more information. Specifically: * now the RX rate information and some of the packet status is passed up; * The 32 bit or 64 bit TSF is passed up; * the PHY mode is passed up; * the "I'm decap'ed AMSDU!" state is passed up; * number of RX chains is bumped to 4. This is all mostly a placeholder for getting the data into the RX status before we pass it up to net80211 - unfortunately we don't yet enforce that drivers provide it, nor do we pass the provided info back up the stack so anyone can use the data. We're going to need to use some of this data moving forward. Notably, now that some hardware can do AMSDU decap for us (the intel iwm driver can do it when we flip it on; the ath10k port I'm doing does it for us) then we need to pass it up through the stack so the duplicate RX sequence numbers and crypto/IV details don't cause the packet to be dropped and/or counted against a replay counter. It's also the beginning of being able to do more interesting node accounting in net80211. Specifically, once drivers start populating per-packet rate information, AMPDU information, timestamps, etc, we can start providing histograms of rate-versus-RSSI, account for receive time spent per node and other such interesting things. (Note: I'm also hoping to include ranging and RTT information for future chipset support; and it's likely going to include it in this kind of fashion.)
2016-10-08 01:12:29 +00:00
rxs.c_rssi = rssi * 2;
rxs.c_nf = sc->sc_noise;
net80211: convert all ieee80211_input_mimo*() consumers to ieee80211_add_rx_params() + drop last (ieee80211_rx_stats) parameter Note: there is an additional check for ieee80211_get_rx_params() return value (which does not exist in the original diff). Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D8207
2016-10-12 20:50:13 +00:00
if (ieee80211_add_rx_params(m, &rxs) == 0)
goto fail;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (ieee80211_radiotap_active_vap(vap)) {
struct iwm_rx_radiotap_header *tap = &sc->sc_rxtap;
tap->wr_flags = 0;
if (phy_info->phy_flags & htole16(IWM_PHY_INFO_FLAG_SHPREAMBLE))
tap->wr_flags |= IEEE80211_RADIOTAP_F_SHORTPRE;
tap->wr_chan_freq = htole16(rxs.c_freq);
/* XXX only if ic->ic_curchan->ic_ieee == rxs.c_ieee */
tap->wr_chan_flags = htole16(ic->ic_curchan->ic_flags);
tap->wr_dbm_antsignal = (int8_t)rssi;
tap->wr_dbm_antnoise = (int8_t)sc->sc_noise;
tap->wr_tsft = phy_info->system_timestamp;
switch (phy_info->rate) {
/* CCK rates. */
case 10: tap->wr_rate = 2; break;
case 20: tap->wr_rate = 4; break;
case 55: tap->wr_rate = 11; break;
case 110: tap->wr_rate = 22; break;
/* OFDM rates. */
case 0xd: tap->wr_rate = 12; break;
case 0xf: tap->wr_rate = 18; break;
case 0x5: tap->wr_rate = 24; break;
case 0x7: tap->wr_rate = 36; break;
case 0x9: tap->wr_rate = 48; break;
case 0xb: tap->wr_rate = 72; break;
case 0x1: tap->wr_rate = 96; break;
case 0x3: tap->wr_rate = 108; break;
/* Unknown rate: should not happen. */
default: tap->wr_rate = 0;
}
}
IWM_UNLOCK(sc);
if (ni != NULL) {
IWM_DPRINTF(sc, IWM_DEBUG_RECV, "input m %p\n", m);
net80211: convert all ieee80211_input_mimo*() consumers to ieee80211_add_rx_params() + drop last (ieee80211_rx_stats) parameter Note: there is an additional check for ieee80211_get_rx_params() return value (which does not exist in the original diff). Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D8207
2016-10-12 20:50:13 +00:00
ieee80211_input_mimo(ni, m);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ieee80211_free_node(ni);
} else {
IWM_DPRINTF(sc, IWM_DEBUG_RECV, "inputall m %p\n", m);
net80211: convert all ieee80211_input_mimo*() consumers to ieee80211_add_rx_params() + drop last (ieee80211_rx_stats) parameter Note: there is an additional check for ieee80211_get_rx_params() return value (which does not exist in the original diff). Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D8207
2016-10-12 20:50:13 +00:00
ieee80211_input_mimo_all(ic, m);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
IWM_LOCK(sc);
iwm: add missing 'Rx errors' counter incrementation.
2016-10-09 10:21:28 +00:00
return;
fail: counter_u64_add(ic->ic_ierrors, 1);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
iwm: convert to ieee80211_tx_complete() Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D4755
2016-01-03 10:10:11 +00:00
static int
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_mvm_rx_tx_cmd_single(struct iwm_softc *sc, struct iwm_rx_packet *pkt,
struct iwm_node *in)
{
struct iwm_mvm_tx_resp *tx_resp = (void *)pkt->data;
net80211: ieee80211_ratectl*: switch to reusable KPI Replace various void * / int argument combinations with common structures: - ieee80211_ratectl_tx_status for *_tx_complete(); - ieee80211_ratectl_tx_stats for *_tx_update(); While here, improve amrr_tx_update() for a bit: 1. In case, if receiver is not known (typical for Ralink USB drivers), refresh Tx rate for all nodes on the interface. 2. There was a misuse: - otus(4) sends non-decreasing counters (as originally intended); - but ural(4), rum(4) and run(4) are using 'read & clear' registers to obtain statistics for some period of time (and those 'last period' values are used as arguments for tx_update()). If arguments are not big enough, they are just discarded after the next call. Fix: move counting into *_tx_update() (now otus(4) will zero out all node counters after every tx_update() call) Tested with: - Intel 3945BG (wpi(4)), STA mode. - WUSB54GC (rum(4)), STA / HOSTAP mode. - RTL8188EU (urtwn(4)), STA mode. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D8037
2016-10-02 20:35:55 +00:00
struct ieee80211_ratectl_tx_status *txs = &sc->sc_txs;
iwm: convert to ieee80211_tx_complete() Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D4755
2016-01-03 10:10:11 +00:00
struct ieee80211_node *ni = &in->in_ni;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
int status = le16toh(tx_resp->status.status) & IWM_TX_STATUS_MSK;
KASSERT(tx_resp->frame_count == 1, ("too many frames"));
/* Update rate control statistics. */
[iwm] add extra transmit setup/completion logging so I can see what's going on. I'm seeing 5GHz association work but data not work until the rate drops, so I need way more information about what's being programmed into the transmit descriptors. Tested: * 7260AC, STA mode
2016-04-26 03:24:28 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_XMIT, "%s: status=0x%04x, seq=%d, fc=%d, btc=%d, frts=%d, ff=%d, irate=%08x, wmt=%d\n",
__func__,
(int) le16toh(tx_resp->status.status),
(int) le16toh(tx_resp->status.sequence),
tx_resp->frame_count,
tx_resp->bt_kill_count,
tx_resp->failure_rts,
tx_resp->failure_frame,
le32toh(tx_resp->initial_rate),
(int) le16toh(tx_resp->wireless_media_time));
net80211: ieee80211_ratectl*: switch to reusable KPI Replace various void * / int argument combinations with common structures: - ieee80211_ratectl_tx_status for *_tx_complete(); - ieee80211_ratectl_tx_stats for *_tx_update(); While here, improve amrr_tx_update() for a bit: 1. In case, if receiver is not known (typical for Ralink USB drivers), refresh Tx rate for all nodes on the interface. 2. There was a misuse: - otus(4) sends non-decreasing counters (as originally intended); - but ural(4), rum(4) and run(4) are using 'read & clear' registers to obtain statistics for some period of time (and those 'last period' values are used as arguments for tx_update()). If arguments are not big enough, they are just discarded after the next call. Fix: move counting into *_tx_update() (now otus(4) will zero out all node counters after every tx_update() call) Tested with: - Intel 3945BG (wpi(4)), STA mode. - WUSB54GC (rum(4)), STA / HOSTAP mode. - RTL8188EU (urtwn(4)), STA mode. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D8037
2016-10-02 20:35:55 +00:00
txs->flags = IEEE80211_RATECTL_STATUS_SHORT_RETRY |
IEEE80211_RATECTL_STATUS_LONG_RETRY;
txs->short_retries = tx_resp->failure_rts;
txs->long_retries = tx_resp->failure_frame;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (status != IWM_TX_STATUS_SUCCESS &&
status != IWM_TX_STATUS_DIRECT_DONE) {
net80211: ieee80211_ratectl*: switch to reusable KPI Replace various void * / int argument combinations with common structures: - ieee80211_ratectl_tx_status for *_tx_complete(); - ieee80211_ratectl_tx_stats for *_tx_update(); While here, improve amrr_tx_update() for a bit: 1. In case, if receiver is not known (typical for Ralink USB drivers), refresh Tx rate for all nodes on the interface. 2. There was a misuse: - otus(4) sends non-decreasing counters (as originally intended); - but ural(4), rum(4) and run(4) are using 'read & clear' registers to obtain statistics for some period of time (and those 'last period' values are used as arguments for tx_update()). If arguments are not big enough, they are just discarded after the next call. Fix: move counting into *_tx_update() (now otus(4) will zero out all node counters after every tx_update() call) Tested with: - Intel 3945BG (wpi(4)), STA mode. - WUSB54GC (rum(4)), STA / HOSTAP mode. - RTL8188EU (urtwn(4)), STA mode. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D8037
2016-10-02 20:35:55 +00:00
switch (status) {
case IWM_TX_STATUS_FAIL_SHORT_LIMIT:
txs->status = IEEE80211_RATECTL_TX_FAIL_SHORT;
break;
case IWM_TX_STATUS_FAIL_LONG_LIMIT:
txs->status = IEEE80211_RATECTL_TX_FAIL_LONG;
break;
case IWM_TX_STATUS_FAIL_LIFE_EXPIRE:
txs->status = IEEE80211_RATECTL_TX_FAIL_EXPIRED;
break;
default:
txs->status = IEEE80211_RATECTL_TX_FAIL_UNSPECIFIED;
break;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
} else {
net80211: ieee80211_ratectl*: switch to reusable KPI Replace various void * / int argument combinations with common structures: - ieee80211_ratectl_tx_status for *_tx_complete(); - ieee80211_ratectl_tx_stats for *_tx_update(); While here, improve amrr_tx_update() for a bit: 1. In case, if receiver is not known (typical for Ralink USB drivers), refresh Tx rate for all nodes on the interface. 2. There was a misuse: - otus(4) sends non-decreasing counters (as originally intended); - but ural(4), rum(4) and run(4) are using 'read & clear' registers to obtain statistics for some period of time (and those 'last period' values are used as arguments for tx_update()). If arguments are not big enough, they are just discarded after the next call. Fix: move counting into *_tx_update() (now otus(4) will zero out all node counters after every tx_update() call) Tested with: - Intel 3945BG (wpi(4)), STA mode. - WUSB54GC (rum(4)), STA / HOSTAP mode. - RTL8188EU (urtwn(4)), STA mode. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D8037
2016-10-02 20:35:55 +00:00
txs->status = IEEE80211_RATECTL_TX_SUCCESS;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
net80211: ieee80211_ratectl*: switch to reusable KPI Replace various void * / int argument combinations with common structures: - ieee80211_ratectl_tx_status for *_tx_complete(); - ieee80211_ratectl_tx_stats for *_tx_update(); While here, improve amrr_tx_update() for a bit: 1. In case, if receiver is not known (typical for Ralink USB drivers), refresh Tx rate for all nodes on the interface. 2. There was a misuse: - otus(4) sends non-decreasing counters (as originally intended); - but ural(4), rum(4) and run(4) are using 'read & clear' registers to obtain statistics for some period of time (and those 'last period' values are used as arguments for tx_update()). If arguments are not big enough, they are just discarded after the next call. Fix: move counting into *_tx_update() (now otus(4) will zero out all node counters after every tx_update() call) Tested with: - Intel 3945BG (wpi(4)), STA mode. - WUSB54GC (rum(4)), STA / HOSTAP mode. - RTL8188EU (urtwn(4)), STA mode. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D8037
2016-10-02 20:35:55 +00:00
ieee80211_ratectl_tx_complete(ni, txs);
return (txs->status != IEEE80211_RATECTL_TX_SUCCESS);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
static void
iwm_mvm_rx_tx_cmd(struct iwm_softc *sc,
struct iwm_rx_packet *pkt, struct iwm_rx_data *data)
{
struct iwm_cmd_header *cmd_hdr = &pkt->hdr;
int idx = cmd_hdr->idx;
int qid = cmd_hdr->qid;
struct iwm_tx_ring *ring = &sc->txq[qid];
struct iwm_tx_data *txd = &ring->data[idx];
struct iwm_node *in = txd->in;
iwm: convert to ieee80211_tx_complete() Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D4755
2016-01-03 10:10:11 +00:00
struct mbuf *m = txd->m;
int status;
KASSERT(txd->done == 0, ("txd not done"));
KASSERT(txd->in != NULL, ("txd without node"));
KASSERT(txd->m != NULL, ("txd without mbuf"));
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
bus_dmamap_sync(ring->data_dmat, data->map, BUS_DMASYNC_POSTREAD);
sc->sc_tx_timer = 0;
iwm: convert to ieee80211_tx_complete() Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D4755
2016-01-03 10:10:11 +00:00
status = iwm_mvm_rx_tx_cmd_single(sc, pkt, in);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Unmap and free mbuf. */
bus_dmamap_sync(ring->data_dmat, txd->map, BUS_DMASYNC_POSTWRITE);
bus_dmamap_unload(ring->data_dmat, txd->map);
IWM_DPRINTF(sc, IWM_DEBUG_XMIT,
"free txd %p, in %p\n", txd, txd->in);
txd->done = 1;
txd->m = NULL;
txd->in = NULL;
iwm: convert to ieee80211_tx_complete() Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D4755
2016-01-03 10:10:11 +00:00
ieee80211_tx_complete(&in->in_ni, m, status);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (--ring->queued < IWM_TX_RING_LOMARK) {
sc->qfullmsk &= ~(1 << ring->qid);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
if (sc->qfullmsk == 0) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* Well, we're in interrupt context, but then again
* I guess net80211 does all sorts of stunts in
* interrupt context, so maybe this is no biggie.
*/
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
iwm_start(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
}
}
/*
* transmit side
*/
/*
* Process a "command done" firmware notification. This is where we wakeup
* processes waiting for a synchronous command completion.
* from if_iwn
*/
static void
iwm_cmd_done(struct iwm_softc *sc, struct iwm_rx_packet *pkt)
{
struct iwm_tx_ring *ring = &sc->txq[IWM_MVM_CMD_QUEUE];
struct iwm_tx_data *data;
if (pkt->hdr.qid != IWM_MVM_CMD_QUEUE) {
return; /* Not a command ack. */
}
iwm: add 'command accepted' debug notification (copied from wpi(4)). Now it should be easier to find out which command causes firmware panics when few commands are sent in a short period of time.
2016-08-27 08:34:20 +00:00
/* XXX wide commands? */
IWM_DPRINTF(sc, IWM_DEBUG_CMD,
"cmd notification type 0x%x qid %d idx %d\n",
pkt->hdr.code, pkt->hdr.qid, pkt->hdr.idx);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
data = &ring->data[pkt->hdr.idx];
/* If the command was mapped in an mbuf, free it. */
if (data->m != NULL) {
bus_dmamap_sync(ring->data_dmat, data->map,
BUS_DMASYNC_POSTWRITE);
bus_dmamap_unload(ring->data_dmat, data->map);
m_freem(data->m);
data->m = NULL;
}
wakeup(&ring->desc[pkt->hdr.idx]);
}
#if 0
/*
* necessary only for block ack mode
*/
void
iwm_update_sched(struct iwm_softc *sc, int qid, int idx, uint8_t sta_id,
uint16_t len)
{
struct iwm_agn_scd_bc_tbl *scd_bc_tbl;
uint16_t w_val;
scd_bc_tbl = sc->sched_dma.vaddr;
len += 8; /* magic numbers came naturally from paris */
if (sc->sc_capaflags & IWM_UCODE_TLV_FLAGS_DW_BC_TABLE)
len = roundup(len, 4) / 4;
w_val = htole16(sta_id << 12 | len);
/* Update TX scheduler. */
scd_bc_tbl[qid].tfd_offset[idx] = w_val;
bus_dmamap_sync(sc->sched_dma.tag, sc->sched_dma.map,
BUS_DMASYNC_PREWRITE);
/* I really wonder what this is ?!? */
if (idx < IWM_TFD_QUEUE_SIZE_BC_DUP) {
scd_bc_tbl[qid].tfd_offset[IWM_TFD_QUEUE_SIZE_MAX + idx] = w_val;
bus_dmamap_sync(sc->sched_dma.tag, sc->sched_dma.map,
BUS_DMASYNC_PREWRITE);
}
}
#endif
/*
* Take an 802.11 (non-n) rate, find the relevant rate
* table entry. return the index into in_ridx[].
*
* The caller then uses that index back into in_ridx
* to figure out the rate index programmed /into/
* the firmware for this given node.
*/
static int
iwm_tx_rateidx_lookup(struct iwm_softc *sc, struct iwm_node *in,
uint8_t rate)
{
int i;
uint8_t r;
for (i = 0; i < nitems(in->in_ridx); i++) {
r = iwm_rates[in->in_ridx[i]].rate;
if (rate == r)
return (i);
}
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_XMIT | IWM_DEBUG_TXRATE,
"%s: couldn't find an entry for rate=%d\n",
__func__,
rate);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* XXX Return the first */
/* XXX TODO: have it return the /lowest/ */
return (0);
}
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
static int
iwm_tx_rateidx_global_lookup(struct iwm_softc *sc, uint8_t rate)
{
int i;
for (i = 0; i < nitems(iwm_rates); i++) {
if (iwm_rates[i].rate == rate)
return (i);
}
/* XXX error? */
IWM_DPRINTF(sc, IWM_DEBUG_XMIT | IWM_DEBUG_TXRATE,
"%s: couldn't find an entry for rate=%d\n",
__func__,
rate);
return (0);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
[iwn] fill in the TX command rate info flags consistently. The Linux driver sets the rate_n_flags regardless of whether it's being sent using firmware rate control or local rate control. This includes the antenna configuration. Thanks to Kyle Evans <kevans91@ksu.edu> for pointing this out to me and doing some investigation/testing on his end. Tested: * Intel 7260 STA, 2G and 5G networks
2016-05-01 05:01:10 +00:00
* Fill in the rate related information for a transmit command.
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
*/
static const struct iwm_rate *
iwm_tx_fill_cmd(struct iwm_softc *sc, struct iwm_node *in,
[iwm] use rate control info from the node txrates; use mgmtrate for EAPOL frames This changes the transmit rate control code to do a few things: * use fixed rates (mcast, ucast, mgmt) where required. * Don't use a hard-coded 11a or 11bg rate for non-data frames - use what net80211 says we should use. * use mgmtrate for EAPOL frames. Reviewed by: avos Differential Revision: https://reviews.freebsd.org/D7994
2016-09-21 20:56:10 +00:00
struct mbuf *m, struct iwm_tx_cmd *tx)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
struct ieee80211_node *ni = &in->in_ni;
[iwm] use rate control info from the node txrates; use mgmtrate for EAPOL frames This changes the transmit rate control code to do a few things: * use fixed rates (mcast, ucast, mgmt) where required. * Don't use a hard-coded 11a or 11bg rate for non-data frames - use what net80211 says we should use. * use mgmtrate for EAPOL frames. Reviewed by: avos Differential Revision: https://reviews.freebsd.org/D7994
2016-09-21 20:56:10 +00:00
struct ieee80211_frame *wh;
const struct ieee80211_txparam *tp = ni->ni_txparms;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
const struct iwm_rate *rinfo;
[iwm] use rate control info from the node txrates; use mgmtrate for EAPOL frames This changes the transmit rate control code to do a few things: * use fixed rates (mcast, ucast, mgmt) where required. * Don't use a hard-coded 11a or 11bg rate for non-data frames - use what net80211 says we should use. * use mgmtrate for EAPOL frames. Reviewed by: avos Differential Revision: https://reviews.freebsd.org/D7994
2016-09-21 20:56:10 +00:00
int type;
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
int ridx, rate_flags;
[iwm] use rate control info from the node txrates; use mgmtrate for EAPOL frames This changes the transmit rate control code to do a few things: * use fixed rates (mcast, ucast, mgmt) where required. * Don't use a hard-coded 11a or 11bg rate for non-data frames - use what net80211 says we should use. * use mgmtrate for EAPOL frames. Reviewed by: avos Differential Revision: https://reviews.freebsd.org/D7994
2016-09-21 20:56:10 +00:00
wh = mtod(m, struct ieee80211_frame *);
type = wh->i_fc[0] & IEEE80211_FC0_TYPE_MASK;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
tx->rts_retry_limit = IWM_RTS_DFAULT_RETRY_LIMIT;
tx->data_retry_limit = IWM_DEFAULT_TX_RETRY;
[iwm] use rate control info from the node txrates; use mgmtrate for EAPOL frames This changes the transmit rate control code to do a few things: * use fixed rates (mcast, ucast, mgmt) where required. * Don't use a hard-coded 11a or 11bg rate for non-data frames - use what net80211 says we should use. * use mgmtrate for EAPOL frames. Reviewed by: avos Differential Revision: https://reviews.freebsd.org/D7994
2016-09-21 20:56:10 +00:00
if (type == IEEE80211_FC0_TYPE_MGT) {
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
ridx = iwm_tx_rateidx_global_lookup(sc, tp->mgmtrate);
IWM_DPRINTF(sc, IWM_DEBUG_TXRATE,
"%s: MGT (%d)\n", __func__, tp->mgmtrate);
[iwm] use rate control info from the node txrates; use mgmtrate for EAPOL frames This changes the transmit rate control code to do a few things: * use fixed rates (mcast, ucast, mgmt) where required. * Don't use a hard-coded 11a or 11bg rate for non-data frames - use what net80211 says we should use. * use mgmtrate for EAPOL frames. Reviewed by: avos Differential Revision: https://reviews.freebsd.org/D7994
2016-09-21 20:56:10 +00:00
} else if (IEEE80211_IS_MULTICAST(wh->i_addr1)) {
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
ridx = iwm_tx_rateidx_global_lookup(sc, tp->mcastrate);
IWM_DPRINTF(sc, IWM_DEBUG_TXRATE,
"%s: MCAST (%d)\n", __func__, tp->mcastrate);
[iwm] use rate control info from the node txrates; use mgmtrate for EAPOL frames This changes the transmit rate control code to do a few things: * use fixed rates (mcast, ucast, mgmt) where required. * Don't use a hard-coded 11a or 11bg rate for non-data frames - use what net80211 says we should use. * use mgmtrate for EAPOL frames. Reviewed by: avos Differential Revision: https://reviews.freebsd.org/D7994
2016-09-21 20:56:10 +00:00
} else if (tp->ucastrate != IEEE80211_FIXED_RATE_NONE) {
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
ridx = iwm_tx_rateidx_global_lookup(sc, tp->ucastrate);
IWM_DPRINTF(sc, IWM_DEBUG_TXRATE,
"%s: FIXED_RATE (%d)\n", __func__, tp->ucastrate);
[iwm] use rate control info from the node txrates; use mgmtrate for EAPOL frames This changes the transmit rate control code to do a few things: * use fixed rates (mcast, ucast, mgmt) where required. * Don't use a hard-coded 11a or 11bg rate for non-data frames - use what net80211 says we should use. * use mgmtrate for EAPOL frames. Reviewed by: avos Differential Revision: https://reviews.freebsd.org/D7994
2016-09-21 20:56:10 +00:00
} else if (m->m_flags & M_EAPOL) {
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
ridx = iwm_tx_rateidx_global_lookup(sc, tp->mgmtrate);
IWM_DPRINTF(sc, IWM_DEBUG_TXRATE,
"%s: EAPOL\n", __func__);
} else if (type == IEEE80211_FC0_TYPE_DATA) {
int i;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* for data frames, use RS table */
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_TXRATE, "%s: DATA\n", __func__);
[iwm] use rate control info from the node txrates; use mgmtrate for EAPOL frames This changes the transmit rate control code to do a few things: * use fixed rates (mcast, ucast, mgmt) where required. * Don't use a hard-coded 11a or 11bg rate for non-data frames - use what net80211 says we should use. * use mgmtrate for EAPOL frames. Reviewed by: avos Differential Revision: https://reviews.freebsd.org/D7994
2016-09-21 20:56:10 +00:00
/* XXX pass pktlen */
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
(void) ieee80211_ratectl_rate(ni, NULL, 0);
i = iwm_tx_rateidx_lookup(sc, in, ni->ni_txrate);
ridx = in->in_ridx[i];
/* This is the index into the programmed table */
tx->initial_rate_index = i;
tx->tx_flags |= htole32(IWM_TX_CMD_FLG_STA_RATE);
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_XMIT | IWM_DEBUG_TXRATE,
"%s: start with i=%d, txrate %d\n",
__func__, i, iwm_rates[ridx].rate);
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
} else {
ridx = iwm_tx_rateidx_global_lookup(sc, tp->mgmtrate);
IWM_DPRINTF(sc, IWM_DEBUG_TXRATE, "%s: DEFAULT (%d)\n",
__func__, tp->mgmtrate);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_XMIT | IWM_DEBUG_TXRATE,
"%s: frame type=%d txrate %d\n",
__func__, type, iwm_rates[ridx].rate);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
rinfo = &iwm_rates[ridx];
IWM_DPRINTF(sc, IWM_DEBUG_TXRATE, "%s: ridx=%d; rate=%d, CCK=%d\n",
__func__, ridx,
rinfo->rate,
!! (IWM_RIDX_IS_CCK(ridx))
);
/* XXX TODO: hard-coded TX antenna? */
rate_flags = 1 << IWM_RATE_MCS_ANT_POS;
if (IWM_RIDX_IS_CCK(ridx))
rate_flags |= IWM_RATE_MCS_CCK_MSK;
tx->rate_n_flags = htole32(rate_flags | rinfo->plcp);
return rinfo;
}
#define TB0_SIZE 16
static int
iwm_tx(struct iwm_softc *sc, struct mbuf *m, struct ieee80211_node *ni, int ac)
{
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct ieee80211com *ic = &sc->sc_ic;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
struct ieee80211vap *vap = TAILQ_FIRST(&ic->ic_vaps);
iwm: reorganize if_iwmvar.h - Change order of data in if_iwmvar.h (like it is in other drivers: defines, data structures, vap/node structures, softc struct and locks); use indentation. - Fix IWM_LOCK(_sc) / IWM_UNLOCK(_sc) macro. - Add IWM_LOCK_INIT / DESTROY(sc) + fix mtx_init() usage. - Wrap iwm_node casts into IWM_NODE() macro. - Drop some fields: * wt_hwqueue from Tx radiotap header; * macaddr[6] from iwm_vap; Approved by: adrian Differential Revision: https://reviews.freebsd.org/D4753
2016-01-03 10:06:10 +00:00
struct iwm_node *in = IWM_NODE(ni);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
struct iwm_tx_ring *ring;
struct iwm_tx_data *data;
struct iwm_tfd *desc;
struct iwm_device_cmd *cmd;
struct iwm_tx_cmd *tx;
struct ieee80211_frame *wh;
struct ieee80211_key *k = NULL;
struct mbuf *m1;
const struct iwm_rate *rinfo;
uint32_t flags;
u_int hdrlen;
bus_dma_segment_t *seg, segs[IWM_MAX_SCATTER];
int nsegs;
uint8_t tid, type;
int i, totlen, error, pad;
wh = mtod(m, struct ieee80211_frame *);
hdrlen = ieee80211_anyhdrsize(wh);
type = wh->i_fc[0] & IEEE80211_FC0_TYPE_MASK;
tid = 0;
ring = &sc->txq[ac];
desc = &ring->desc[ring->cur];
memset(desc, 0, sizeof(*desc));
data = &ring->data[ring->cur];
/* Fill out iwm_tx_cmd to send to the firmware */
cmd = &ring->cmd[ring->cur];
cmd->hdr.code = IWM_TX_CMD;
cmd->hdr.flags = 0;
cmd->hdr.qid = ring->qid;
cmd->hdr.idx = ring->cur;
tx = (void *)cmd->data;
memset(tx, 0, sizeof(*tx));
[iwm] use rate control info from the node txrates; use mgmtrate for EAPOL frames This changes the transmit rate control code to do a few things: * use fixed rates (mcast, ucast, mgmt) where required. * Don't use a hard-coded 11a or 11bg rate for non-data frames - use what net80211 says we should use. * use mgmtrate for EAPOL frames. Reviewed by: avos Differential Revision: https://reviews.freebsd.org/D7994
2016-09-21 20:56:10 +00:00
rinfo = iwm_tx_fill_cmd(sc, in, m, tx);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Encrypt the frame if need be. */
if (wh->i_fc[1] & IEEE80211_FC1_PROTECTED) {
/* Retrieve key for TX && do software encryption. */
k = ieee80211_crypto_encap(ni, m);
if (k == NULL) {
m_freem(m);
return (ENOBUFS);
}
/* 802.11 header may have moved. */
wh = mtod(m, struct ieee80211_frame *);
}
if (ieee80211_radiotap_active_vap(vap)) {
struct iwm_tx_radiotap_header *tap = &sc->sc_txtap;
tap->wt_flags = 0;
tap->wt_chan_freq = htole16(ni->ni_chan->ic_freq);
tap->wt_chan_flags = htole16(ni->ni_chan->ic_flags);
tap->wt_rate = rinfo->rate;
if (k != NULL)
tap->wt_flags |= IEEE80211_RADIOTAP_F_WEP;
ieee80211_radiotap_tx(vap, m);
}
totlen = m->m_pkthdr.len;
flags = 0;
if (!IEEE80211_IS_MULTICAST(wh->i_addr1)) {
flags |= IWM_TX_CMD_FLG_ACK;
}
[iwm] Fix inverted logic in iwm_tx(). The PROT_REQUIRE flag in should be set for data frames above a certain length, but we were setting it for !data frames above a certain length, which makes no sense at all. Taken-From: OpenBSD, Linux iwlwifi Approved by: adrian (mentor) Obtained from: DragonFlyBSD git 8cc03924a36c572c2908e659e624f44636dc2b33 Differential Revision: https://reviews.freebsd.org/D7323
2016-07-27 20:43:08 +00:00
if (type == IEEE80211_FC0_TYPE_DATA
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
&& (totlen + IEEE80211_CRC_LEN > vap->iv_rtsthreshold)
&& !IEEE80211_IS_MULTICAST(wh->i_addr1)) {
flags |= IWM_TX_CMD_FLG_PROT_REQUIRE;
}
if (IEEE80211_IS_MULTICAST(wh->i_addr1) ||
type != IEEE80211_FC0_TYPE_DATA)
tx->sta_id = sc->sc_aux_sta.sta_id;
else
tx->sta_id = IWM_STATION_ID;
if (type == IEEE80211_FC0_TYPE_MGT) {
uint8_t subtype = wh->i_fc[0] & IEEE80211_FC0_SUBTYPE_MASK;
if (subtype == IEEE80211_FC0_SUBTYPE_ASSOC_REQ ||
[iwm] Set different pm_timeout for action frames. When building a Tx Command for management frames, we are lacking a check for action frames, for which we should set a different pm_timeout. This cause the fw to stay awake for 100TU after each such frame is transmitted, resulting an excessive power consumption. Taken-From: Linux iwlwifi (git b084a35663c3f1f7) Approved by: adrian (mentor) Obtained from: Linux git b084a35663c3f1f7de1c45c4ae3006864c940fe7 Obtained from: DragonFlyBSD git ba00f0e3ae873d6f0d5743e22c3ebc49c44dfdac Differential Revision: https://reviews.freebsd.org/D7324
2016-07-27 20:46:51 +00:00
subtype == IEEE80211_FC0_SUBTYPE_REASSOC_REQ) {
tx->pm_frame_timeout = htole16(IWM_PM_FRAME_ASSOC);
} else if (subtype == IEEE80211_FC0_SUBTYPE_ACTION) {
tx->pm_frame_timeout = htole16(IWM_PM_FRAME_NONE);
} else {
tx->pm_frame_timeout = htole16(IWM_PM_FRAME_MGMT);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
} else {
[iwm] Set different pm_timeout for action frames. When building a Tx Command for management frames, we are lacking a check for action frames, for which we should set a different pm_timeout. This cause the fw to stay awake for 100TU after each such frame is transmitted, resulting an excessive power consumption. Taken-From: Linux iwlwifi (git b084a35663c3f1f7) Approved by: adrian (mentor) Obtained from: Linux git b084a35663c3f1f7de1c45c4ae3006864c940fe7 Obtained from: DragonFlyBSD git ba00f0e3ae873d6f0d5743e22c3ebc49c44dfdac Differential Revision: https://reviews.freebsd.org/D7324
2016-07-27 20:46:51 +00:00
tx->pm_frame_timeout = htole16(IWM_PM_FRAME_NONE);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
if (hdrlen & 3) {
/* First segment length must be a multiple of 4. */
flags |= IWM_TX_CMD_FLG_MH_PAD;
pad = 4 - (hdrlen & 3);
} else
pad = 0;
tx->driver_txop = 0;
tx->next_frame_len = 0;
tx->len = htole16(totlen);
tx->tid_tspec = tid;
tx->life_time = htole32(IWM_TX_CMD_LIFE_TIME_INFINITE);
/* Set physical address of "scratch area". */
tx->dram_lsb_ptr = htole32(data->scratch_paddr);
tx->dram_msb_ptr = iwm_get_dma_hi_addr(data->scratch_paddr);
/* Copy 802.11 header in TX command. */
memcpy(((uint8_t *)tx) + sizeof(*tx), wh, hdrlen);
flags |= IWM_TX_CMD_FLG_BT_DIS | IWM_TX_CMD_FLG_SEQ_CTL;
tx->sec_ctl = 0;
tx->tx_flags |= htole32(flags);
/* Trim 802.11 header. */
m_adj(m, hdrlen);
error = bus_dmamap_load_mbuf_sg(ring->data_dmat, data->map, m,
segs, &nsegs, BUS_DMA_NOWAIT);
if (error != 0) {
if (error != EFBIG) {
device_printf(sc->sc_dev, "can't map mbuf (error %d)\n",
error);
m_freem(m);
return error;
}
/* Too many DMA segments, linearize mbuf. */
iwm: use m_collapse() to defragment a mbuf chain - Simplify defragmentation code. - Use proper number of dma segments for data. Approved by: adrian (mentor) Obtained from: DragonFlyBSD (mostly) Differential Revision: https://reviews.freebsd.org/D4754
2016-01-03 21:32:47 +00:00
m1 = m_collapse(m, M_NOWAIT, IWM_MAX_SCATTER - 2);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (m1 == NULL) {
iwm: use m_collapse() to defragment a mbuf chain - Simplify defragmentation code. - Use proper number of dma segments for data. Approved by: adrian (mentor) Obtained from: DragonFlyBSD (mostly) Differential Revision: https://reviews.freebsd.org/D4754
2016-01-03 21:32:47 +00:00
device_printf(sc->sc_dev,
"%s: could not defrag mbuf\n", __func__);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
m_freem(m);
iwm: use m_collapse() to defragment a mbuf chain - Simplify defragmentation code. - Use proper number of dma segments for data. Approved by: adrian (mentor) Obtained from: DragonFlyBSD (mostly) Differential Revision: https://reviews.freebsd.org/D4754
2016-01-03 21:32:47 +00:00
return (ENOBUFS);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
m = m1;
iwm: use m_collapse() to defragment a mbuf chain - Simplify defragmentation code. - Use proper number of dma segments for data. Approved by: adrian (mentor) Obtained from: DragonFlyBSD (mostly) Differential Revision: https://reviews.freebsd.org/D4754
2016-01-03 21:32:47 +00:00
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
error = bus_dmamap_load_mbuf_sg(ring->data_dmat, data->map, m,
segs, &nsegs, BUS_DMA_NOWAIT);
if (error != 0) {
device_printf(sc->sc_dev, "can't map mbuf (error %d)\n",
error);
m_freem(m);
return error;
}
}
data->m = m;
data->in = in;
data->done = 0;
IWM_DPRINTF(sc, IWM_DEBUG_XMIT,
"sending txd %p, in %p\n", data, data->in);
KASSERT(data->in != NULL, ("node is NULL"));
IWM_DPRINTF(sc, IWM_DEBUG_XMIT,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
"sending data: qid=%d idx=%d len=%d nsegs=%d txflags=0x%08x rate_n_flags=0x%08x rateidx=%u\n",
[iwm] add extra transmit setup/completion logging so I can see what's going on. I'm seeing 5GHz association work but data not work until the rate drops, so I need way more information about what's being programmed into the transmit descriptors. Tested: * 7260AC, STA mode
2016-04-26 03:24:28 +00:00
ring->qid, ring->cur, totlen, nsegs,
le32toh(tx->tx_flags),
le32toh(tx->rate_n_flags),
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
tx->initial_rate_index
[iwm] add extra transmit setup/completion logging so I can see what's going on. I'm seeing 5GHz association work but data not work until the rate drops, so I need way more information about what's being programmed into the transmit descriptors. Tested: * 7260AC, STA mode
2016-04-26 03:24:28 +00:00
);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Fill TX descriptor. */
desc->num_tbs = 2 + nsegs;
desc->tbs[0].lo = htole32(data->cmd_paddr);
desc->tbs[0].hi_n_len = htole16(iwm_get_dma_hi_addr(data->cmd_paddr)) |
(TB0_SIZE << 4);
desc->tbs[1].lo = htole32(data->cmd_paddr + TB0_SIZE);
desc->tbs[1].hi_n_len = htole16(iwm_get_dma_hi_addr(data->cmd_paddr)) |
((sizeof(struct iwm_cmd_header) + sizeof(*tx)
+ hdrlen + pad - TB0_SIZE) << 4);
/* Other DMA segments are for data payload. */
for (i = 0; i < nsegs; i++) {
seg = &segs[i];
desc->tbs[i+2].lo = htole32(seg->ds_addr);
desc->tbs[i+2].hi_n_len = \
htole16(iwm_get_dma_hi_addr(seg->ds_addr))
| ((seg->ds_len) << 4);
}
bus_dmamap_sync(ring->data_dmat, data->map,
BUS_DMASYNC_PREWRITE);
bus_dmamap_sync(ring->cmd_dma.tag, ring->cmd_dma.map,
BUS_DMASYNC_PREWRITE);
bus_dmamap_sync(ring->desc_dma.tag, ring->desc_dma.map,
BUS_DMASYNC_PREWRITE);
#if 0
iwm_update_sched(sc, ring->qid, ring->cur, tx->sta_id, le16toh(tx->len));
#endif
/* Kick TX ring. */
ring->cur = (ring->cur + 1) % IWM_TX_RING_COUNT;
IWM_WRITE(sc, IWM_HBUS_TARG_WRPTR, ring->qid << 8 | ring->cur);
/* Mark TX ring as full if we reach a certain threshold. */
if (++ring->queued > IWM_TX_RING_HIMARK) {
sc->qfullmsk |= 1 << ring->qid;
}
return 0;
}
static int
iwm_raw_xmit(struct ieee80211_node *ni, struct mbuf *m,
const struct ieee80211_bpf_params *params)
{
struct ieee80211com *ic = ni->ni_ic;
Convert more ifp->if_softc -> ic->ic_softc. These should be a big no-op. Tested: * make universe
2015-08-17 23:44:38 +00:00
struct iwm_softc *sc = ic->ic_softc;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
int error = 0;
IWM_DPRINTF(sc, IWM_DEBUG_XMIT,
"->%s begin\n", __func__);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
if ((sc->sc_flags & IWM_FLAG_HW_INITED) == 0) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
m_freem(m);
IWM_DPRINTF(sc, IWM_DEBUG_XMIT,
"<-%s not RUNNING\n", __func__);
return (ENETDOWN);
}
IWM_LOCK(sc);
/* XXX fix this */
if (params == NULL) {
error = iwm_tx(sc, m, ni, 0);
} else {
error = iwm_tx(sc, m, ni, 0);
}
sc->sc_tx_timer = 5;
IWM_UNLOCK(sc);
return (error);
}
/*
* mvm/tx.c
*/
/*
* Note that there are transports that buffer frames before they reach
* the firmware. This means that after flush_tx_path is called, the
* queue might not be empty. The race-free way to handle this is to:
* 1) set the station as draining
* 2) flush the Tx path
* 3) wait for the transport queues to be empty
*/
int
[iwm] Improve reliability of iwm_release on disassociation a bit. * We need to first call ivp->iv_newstate(), to enqueue the deauth/deassoc mgmt frame, then flush the tx queue, before actually calling iwm_release(). * cycling a wlan connection via wpa_cli frontend to wpa_supplicant, by issuing disconnect and reconnect commandos works quite well. (There is still an issue when disconnecting/reconnecting too quickly) * Reassociating or roaming via wpa_supplicant is still broken. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7943
2016-09-19 22:36:30 +00:00
iwm_mvm_flush_tx_path(struct iwm_softc *sc, uint32_t tfd_msk, uint32_t flags)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
[iwm] Improve reliability of iwm_release on disassociation a bit. * We need to first call ivp->iv_newstate(), to enqueue the deauth/deassoc mgmt frame, then flush the tx queue, before actually calling iwm_release(). * cycling a wlan connection via wpa_cli frontend to wpa_supplicant, by issuing disconnect and reconnect commandos works quite well. (There is still an issue when disconnecting/reconnecting too quickly) * Reassociating or roaming via wpa_supplicant is still broken. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7943
2016-09-19 22:36:30 +00:00
int ret;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
struct iwm_tx_path_flush_cmd flush_cmd = {
.queues_ctl = htole32(tfd_msk),
.flush_ctl = htole16(IWM_DUMP_TX_FIFO_FLUSH),
};
[iwm] Improve reliability of iwm_release on disassociation a bit. * We need to first call ivp->iv_newstate(), to enqueue the deauth/deassoc mgmt frame, then flush the tx queue, before actually calling iwm_release(). * cycling a wlan connection via wpa_cli frontend to wpa_supplicant, by issuing disconnect and reconnect commandos works quite well. (There is still an issue when disconnecting/reconnecting too quickly) * Reassociating or roaming via wpa_supplicant is still broken. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7943
2016-09-19 22:36:30 +00:00
ret = iwm_mvm_send_cmd_pdu(sc, IWM_TXPATH_FLUSH, flags,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
sizeof(flush_cmd), &flush_cmd);
if (ret)
device_printf(sc->sc_dev,
"Flushing tx queue failed: %d\n", ret);
return ret;
}
/*
* BEGIN mvm/sta.c
*/
static int
iwm_mvm_send_add_sta_cmd_status(struct iwm_softc *sc,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
struct iwm_mvm_add_sta_cmd_v7 *cmd, int *status)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
return iwm_mvm_send_cmd_pdu_status(sc, IWM_ADD_STA, sizeof(*cmd),
cmd, status);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
/* send station add/update command to firmware */
static int
iwm_mvm_sta_send_to_fw(struct iwm_softc *sc, struct iwm_node *in, int update)
{
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
struct iwm_mvm_add_sta_cmd_v7 add_sta_cmd;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
int ret;
uint32_t status;
memset(&add_sta_cmd, 0, sizeof(add_sta_cmd));
add_sta_cmd.sta_id = IWM_STATION_ID;
add_sta_cmd.mac_id_n_color
= htole32(IWM_FW_CMD_ID_AND_COLOR(IWM_DEFAULT_MACID,
IWM_DEFAULT_COLOR));
if (!update) {
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
int ac;
for (ac = 0; ac < WME_NUM_AC; ac++) {
add_sta_cmd.tfd_queue_msk |=
htole32(1 << iwm_mvm_ac_to_tx_fifo[ac]);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IEEE80211_ADDR_COPY(&add_sta_cmd.addr, in->in_ni.ni_bssid);
}
add_sta_cmd.add_modify = update ? 1 : 0;
add_sta_cmd.station_flags_msk
|= htole32(IWM_STA_FLG_FAT_EN_MSK | IWM_STA_FLG_MIMO_EN_MSK);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
add_sta_cmd.tid_disable_tx = htole16(0xffff);
if (update)
add_sta_cmd.modify_mask |= (IWM_STA_MODIFY_TID_DISABLE_TX);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
status = IWM_ADD_STA_SUCCESS;
ret = iwm_mvm_send_add_sta_cmd_status(sc, &add_sta_cmd, &status);
if (ret)
return ret;
switch (status) {
case IWM_ADD_STA_SUCCESS:
break;
default:
ret = EIO;
device_printf(sc->sc_dev, "IWM_ADD_STA failed\n");
break;
}
return ret;
}
static int
iwm_mvm_add_sta(struct iwm_softc *sc, struct iwm_node *in)
{
[iwm] Fix several nitpicks in iwm(4). Move some declarations to if_iwmreg.h. Remove iwm_fw_alive(); just call iwm_post_alive() directly. Simplify iwm_mvm_add_sta(). Return timeout error from iwm_apm_init(). Print a message when init (i.e. boot) firmware fails to load. Remove some commented-out code which wouldn't compile anyway. Move iwm_mvm_tx_fifo to if_iwmreg.h to match better where Linux puts it. Taken-From: OpenBSD (if_iwm.c r1.80 and if_iwmreg.h r1.11) Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD 29fcb331e5620ae145a6ab9cdda830e22fff626a
2016-06-02 05:43:16 +00:00
return iwm_mvm_sta_send_to_fw(sc, in, 0);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
static int
iwm_mvm_update_sta(struct iwm_softc *sc, struct iwm_node *in)
{
return iwm_mvm_sta_send_to_fw(sc, in, 1);
}
static int
iwm_mvm_add_int_sta_common(struct iwm_softc *sc, struct iwm_int_sta *sta,
const uint8_t *addr, uint16_t mac_id, uint16_t color)
{
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
struct iwm_mvm_add_sta_cmd_v7 cmd;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
int ret;
uint32_t status;
memset(&cmd, 0, sizeof(cmd));
cmd.sta_id = sta->sta_id;
cmd.mac_id_n_color = htole32(IWM_FW_CMD_ID_AND_COLOR(mac_id, color));
cmd.tfd_queue_msk = htole32(sta->tfd_queue_msk);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
cmd.tid_disable_tx = htole16(0xffff);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (addr)
IEEE80211_ADDR_COPY(cmd.addr, addr);
ret = iwm_mvm_send_add_sta_cmd_status(sc, &cmd, &status);
if (ret)
return ret;
switch (status) {
case IWM_ADD_STA_SUCCESS:
IWM_DPRINTF(sc, IWM_DEBUG_RESET,
"%s: Internal station added.\n", __func__);
return 0;
default:
device_printf(sc->sc_dev,
"%s: Add internal station failed, status=0x%x\n",
__func__, status);
ret = EIO;
break;
}
return ret;
}
static int
iwm_mvm_add_aux_sta(struct iwm_softc *sc)
{
int ret;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sc->sc_aux_sta.sta_id = IWM_AUX_STA_ID;
sc->sc_aux_sta.tfd_queue_msk = (1 << IWM_MVM_AUX_QUEUE);
ret = iwm_enable_txq(sc, 0, IWM_MVM_AUX_QUEUE, IWM_MVM_TX_FIFO_MCAST);
if (ret)
return ret;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ret = iwm_mvm_add_int_sta_common(sc,
&sc->sc_aux_sta, NULL, IWM_MAC_INDEX_AUX, 0);
if (ret)
memset(&sc->sc_aux_sta, 0, sizeof(sc->sc_aux_sta));
return ret;
}
/*
* END mvm/sta.c
*/
/*
* BEGIN mvm/quota.c
*/
static int
iwm_mvm_update_quotas(struct iwm_softc *sc, struct iwm_node *in)
{
struct iwm_time_quota_cmd cmd;
int i, idx, ret, num_active_macs, quota, quota_rem;
int colors[IWM_MAX_BINDINGS] = { -1, -1, -1, -1, };
int n_ifs[IWM_MAX_BINDINGS] = {0, };
uint16_t id;
memset(&cmd, 0, sizeof(cmd));
/* currently, PHY ID == binding ID */
if (in) {
id = in->in_phyctxt->id;
KASSERT(id < IWM_MAX_BINDINGS, ("invalid id"));
colors[id] = in->in_phyctxt->color;
if (1)
n_ifs[id] = 1;
}
/*
* The FW's scheduling session consists of
* IWM_MVM_MAX_QUOTA fragments. Divide these fragments
* equally between all the bindings that require quota
*/
num_active_macs = 0;
for (i = 0; i < IWM_MAX_BINDINGS; i++) {
cmd.quotas[i].id_and_color = htole32(IWM_FW_CTXT_INVALID);
num_active_macs += n_ifs[i];
}
quota = 0;
quota_rem = 0;
if (num_active_macs) {
quota = IWM_MVM_MAX_QUOTA / num_active_macs;
quota_rem = IWM_MVM_MAX_QUOTA % num_active_macs;
}
for (idx = 0, i = 0; i < IWM_MAX_BINDINGS; i++) {
if (colors[i] < 0)
continue;
cmd.quotas[idx].id_and_color =
htole32(IWM_FW_CMD_ID_AND_COLOR(i, colors[i]));
if (n_ifs[i] <= 0) {
cmd.quotas[idx].quota = htole32(0);
cmd.quotas[idx].max_duration = htole32(0);
} else {
cmd.quotas[idx].quota = htole32(quota * n_ifs[i]);
cmd.quotas[idx].max_duration = htole32(0);
}
idx++;
}
/* Give the remainder of the session to the first binding */
cmd.quotas[0].quota = htole32(le32toh(cmd.quotas[0].quota) + quota_rem);
ret = iwm_mvm_send_cmd_pdu(sc, IWM_TIME_QUOTA_CMD, IWM_CMD_SYNC,
sizeof(cmd), &cmd);
if (ret)
device_printf(sc->sc_dev,
"%s: Failed to send quota: %d\n", __func__, ret);
return ret;
}
/*
* END mvm/quota.c
*/
/*
* ieee80211 routines
*/
/*
* Change to AUTH state in 80211 state machine. Roughly matches what
* Linux does in bss_info_changed().
*/
static int
iwm_auth(struct ieee80211vap *vap, struct iwm_softc *sc)
{
struct ieee80211_node *ni;
struct iwm_node *in;
struct iwm_vap *iv = IWM_VAP(vap);
uint32_t duration;
int error;
/*
* XXX i have a feeling that the vap node is being
* freed from underneath us. Grr.
*/
ni = ieee80211_ref_node(vap->iv_bss);
iwm: reorganize if_iwmvar.h - Change order of data in if_iwmvar.h (like it is in other drivers: defines, data structures, vap/node structures, softc struct and locks); use indentation. - Fix IWM_LOCK(_sc) / IWM_UNLOCK(_sc) macro. - Add IWM_LOCK_INIT / DESTROY(sc) + fix mtx_init() usage. - Wrap iwm_node casts into IWM_NODE() macro. - Drop some fields: * wt_hwqueue from Tx radiotap header; * macaddr[6] from iwm_vap; Approved by: adrian Differential Revision: https://reviews.freebsd.org/D4753
2016-01-03 10:06:10 +00:00
in = IWM_NODE(ni);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_RESET | IWM_DEBUG_STATE,
"%s: called; vap=%p, bss ni=%p\n",
__func__,
vap,
ni);
in->in_assoc = 0;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
error = iwm_mvm_sf_config(sc, IWM_SF_FULL_ON);
if (error != 0)
return error;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
error = iwm_allow_mcast(vap, sc);
if (error) {
device_printf(sc->sc_dev,
"%s: failed to set multicast\n", __func__);
goto out;
}
/*
* This is where it deviates from what Linux does.
*
* Linux iwlwifi doesn't reset the nic each time, nor does it
* call ctxt_add() here. Instead, it adds it during vap creation,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
* and always does a mac_ctx_changed().
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
*
* The openbsd port doesn't attempt to do that - it reset things
* at odd states and does the add here.
*
* So, until the state handling is fixed (ie, we never reset
* the NIC except for a firmware failure, which should drag
* the NIC back to IDLE, re-setup and re-add all the mac/phy
* contexts that are required), let's do a dirty hack here.
*/
if (iv->is_uploaded) {
if ((error = iwm_mvm_mac_ctxt_changed(sc, vap)) != 0) {
device_printf(sc->sc_dev,
[iwn] fix firmware command use in iwm_auth(). The iwm firmware has separate commands for add, modify and delete for various things (mac, phy context, etc.) The openbsd driver has a habit of just completely resetting the NIC each time, which is technically mostly okay (as long as the reset doesn't actually fail!) but it means a lot of the code is doing ADD when it should do MODIFY. The firmware responds in kind - it just asserts. This fixes auth attempts that occur after the NIC has been already configured. (I'm sure there are more instances of this!) Tested: iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: revision: 0x140, firmware 25.228 (API ver. 9) .. STA mode. Submitted by: Masachika ISHIZUKA <ish@amail.plala.or.jp>
2016-04-25 16:13:04 +00:00
"%s: failed to update MAC\n", __func__);
goto out;
}
if ((error = iwm_mvm_phy_ctxt_changed(sc, &sc->sc_phyctxt[0],
in->in_ni.ni_chan, 1, 1)) != 0) {
device_printf(sc->sc_dev,
"%s: failed update phy ctxt\n", __func__);
goto out;
}
in->in_phyctxt = &sc->sc_phyctxt[0];
if ((error = iwm_mvm_binding_update(sc, in)) != 0) {
device_printf(sc->sc_dev,
"%s: binding update cmd\n", __func__);
goto out;
}
if ((error = iwm_mvm_update_sta(sc, in)) != 0) {
device_printf(sc->sc_dev,
"%s: failed to update sta\n", __func__);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
goto out;
}
} else {
if ((error = iwm_mvm_mac_ctxt_add(sc, vap)) != 0) {
device_printf(sc->sc_dev,
"%s: failed to add MAC\n", __func__);
goto out;
}
[iwn] fix firmware command use in iwm_auth(). The iwm firmware has separate commands for add, modify and delete for various things (mac, phy context, etc.) The openbsd driver has a habit of just completely resetting the NIC each time, which is technically mostly okay (as long as the reset doesn't actually fail!) but it means a lot of the code is doing ADD when it should do MODIFY. The firmware responds in kind - it just asserts. This fixes auth attempts that occur after the NIC has been already configured. (I'm sure there are more instances of this!) Tested: iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: revision: 0x140, firmware 25.228 (API ver. 9) .. STA mode. Submitted by: Masachika ISHIZUKA <ish@amail.plala.or.jp>
2016-04-25 16:13:04 +00:00
if ((error = iwm_mvm_phy_ctxt_changed(sc, &sc->sc_phyctxt[0],
in->in_ni.ni_chan, 1, 1)) != 0) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev,
[iwn] fix firmware command use in iwm_auth(). The iwm firmware has separate commands for add, modify and delete for various things (mac, phy context, etc.) The openbsd driver has a habit of just completely resetting the NIC each time, which is technically mostly okay (as long as the reset doesn't actually fail!) but it means a lot of the code is doing ADD when it should do MODIFY. The firmware responds in kind - it just asserts. This fixes auth attempts that occur after the NIC has been already configured. (I'm sure there are more instances of this!) Tested: iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: revision: 0x140, firmware 25.228 (API ver. 9) .. STA mode. Submitted by: Masachika ISHIZUKA <ish@amail.plala.or.jp>
2016-04-25 16:13:04 +00:00
"%s: failed add phy ctxt!\n", __func__);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
error = ETIMEDOUT;
goto out;
[iwn] fix firmware command use in iwm_auth(). The iwm firmware has separate commands for add, modify and delete for various things (mac, phy context, etc.) The openbsd driver has a habit of just completely resetting the NIC each time, which is technically mostly okay (as long as the reset doesn't actually fail!) but it means a lot of the code is doing ADD when it should do MODIFY. The firmware responds in kind - it just asserts. This fixes auth attempts that occur after the NIC has been already configured. (I'm sure there are more instances of this!) Tested: iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: revision: 0x140, firmware 25.228 (API ver. 9) .. STA mode. Submitted by: Masachika ISHIZUKA <ish@amail.plala.or.jp>
2016-04-25 16:13:04 +00:00
}
in->in_phyctxt = &sc->sc_phyctxt[0];
if ((error = iwm_mvm_binding_add_vif(sc, in)) != 0) {
device_printf(sc->sc_dev,
"%s: binding add cmd\n", __func__);
goto out;
}
if ((error = iwm_mvm_add_sta(sc, in)) != 0) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev,
[iwn] fix firmware command use in iwm_auth(). The iwm firmware has separate commands for add, modify and delete for various things (mac, phy context, etc.) The openbsd driver has a habit of just completely resetting the NIC each time, which is technically mostly okay (as long as the reset doesn't actually fail!) but it means a lot of the code is doing ADD when it should do MODIFY. The firmware responds in kind - it just asserts. This fixes auth attempts that occur after the NIC has been already configured. (I'm sure there are more instances of this!) Tested: iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: revision: 0x140, firmware 25.228 (API ver. 9) .. STA mode. Submitted by: Masachika ISHIZUKA <ish@amail.plala.or.jp>
2016-04-25 16:13:04 +00:00
"%s: failed to add sta\n", __func__);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
goto out;
}
}
[iwn] fix firmware command use in iwm_auth(). The iwm firmware has separate commands for add, modify and delete for various things (mac, phy context, etc.) The openbsd driver has a habit of just completely resetting the NIC each time, which is technically mostly okay (as long as the reset doesn't actually fail!) but it means a lot of the code is doing ADD when it should do MODIFY. The firmware responds in kind - it just asserts. This fixes auth attempts that occur after the NIC has been already configured. (I'm sure there are more instances of this!) Tested: iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: revision: 0x140, firmware 25.228 (API ver. 9) .. STA mode. Submitted by: Masachika ISHIZUKA <ish@amail.plala.or.jp>
2016-04-25 16:13:04 +00:00
/*
* Prevent the FW from wandering off channel during association
* by "protecting" the session with a time event.
*/
/* XXX duration is in units of TU, not MS */
duration = IWM_MVM_TE_SESSION_PROTECTION_MAX_TIME_MS;
iwm_mvm_protect_session(sc, in, duration, 500 /* XXX magic number */);
DELAY(100);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
error = 0;
out:
ieee80211_free_node(ni);
return (error);
}
static int
iwm_assoc(struct ieee80211vap *vap, struct iwm_softc *sc)
{
iwm: reorganize if_iwmvar.h - Change order of data in if_iwmvar.h (like it is in other drivers: defines, data structures, vap/node structures, softc struct and locks); use indentation. - Fix IWM_LOCK(_sc) / IWM_UNLOCK(_sc) macro. - Add IWM_LOCK_INIT / DESTROY(sc) + fix mtx_init() usage. - Wrap iwm_node casts into IWM_NODE() macro. - Drop some fields: * wt_hwqueue from Tx radiotap header; * macaddr[6] from iwm_vap; Approved by: adrian Differential Revision: https://reviews.freebsd.org/D4753
2016-01-03 10:06:10 +00:00
struct iwm_node *in = IWM_NODE(vap->iv_bss);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
int error;
if ((error = iwm_mvm_update_sta(sc, in)) != 0) {
device_printf(sc->sc_dev,
"%s: failed to update STA\n", __func__);
return error;
}
in->in_assoc = 1;
if ((error = iwm_mvm_mac_ctxt_changed(sc, vap)) != 0) {
device_printf(sc->sc_dev,
"%s: failed to update MAC\n", __func__);
return error;
}
return 0;
}
static int
iwm_release(struct iwm_softc *sc, struct iwm_node *in)
{
[iwm] Improve reliability of iwm_release on disassociation a bit. * We need to first call ivp->iv_newstate(), to enqueue the deauth/deassoc mgmt frame, then flush the tx queue, before actually calling iwm_release(). * cycling a wlan connection via wpa_cli frontend to wpa_supplicant, by issuing disconnect and reconnect commandos works quite well. (There is still an issue when disconnecting/reconnecting too quickly) * Reassociating or roaming via wpa_supplicant is still broken. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7943
2016-09-19 22:36:30 +00:00
uint32_t tfd_msk;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* Ok, so *technically* the proper set of calls for going
* from RUN back to SCAN is:
*
* iwm_mvm_power_mac_disable(sc, in);
* iwm_mvm_mac_ctxt_changed(sc, in);
* iwm_mvm_rm_sta(sc, in);
* iwm_mvm_update_quotas(sc, NULL);
* iwm_mvm_mac_ctxt_changed(sc, in);
* iwm_mvm_binding_remove_vif(sc, in);
* iwm_mvm_mac_ctxt_remove(sc, in);
*
* However, that freezes the device not matter which permutations
* and modifications are attempted. Obviously, this driver is missing
* something since it works in the Linux driver, but figuring out what
* is missing is a little more complicated. Now, since we're going
* back to nothing anyway, we'll just do a complete device reset.
* Up your's, device!
*/
[iwm] Improve reliability of iwm_release on disassociation a bit. * We need to first call ivp->iv_newstate(), to enqueue the deauth/deassoc mgmt frame, then flush the tx queue, before actually calling iwm_release(). * cycling a wlan connection via wpa_cli frontend to wpa_supplicant, by issuing disconnect and reconnect commandos works quite well. (There is still an issue when disconnecting/reconnecting too quickly) * Reassociating or roaming via wpa_supplicant is still broken. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7943
2016-09-19 22:36:30 +00:00
/*
* Just using 0xf for the queues mask is fine as long as we only
* get here from RUN state.
*/
tfd_msk = 0xf;
mbufq_drain(&sc->sc_snd);
iwm_mvm_flush_tx_path(sc, tfd_msk, IWM_CMD_SYNC);
/*
* We seem to get away with just synchronously sending the
* IWM_TXPATH_FLUSH command.
*/
// iwm_trans_wait_tx_queue_empty(sc, tfd_msk);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_stop_device(sc);
iwm_init_hw(sc);
if (in)
in->in_assoc = 0;
return 0;
#if 0
int error;
iwm_mvm_power_mac_disable(sc, in);
if ((error = iwm_mvm_mac_ctxt_changed(sc, in)) != 0) {
device_printf(sc->sc_dev, "mac ctxt change fail 1 %d\n", error);
return error;
}
if ((error = iwm_mvm_rm_sta(sc, in)) != 0) {
device_printf(sc->sc_dev, "sta remove fail %d\n", error);
return error;
}
error = iwm_mvm_rm_sta(sc, in);
in->in_assoc = 0;
iwm_mvm_update_quotas(sc, NULL);
if ((error = iwm_mvm_mac_ctxt_changed(sc, in)) != 0) {
device_printf(sc->sc_dev, "mac ctxt change fail 2 %d\n", error);
return error;
}
iwm_mvm_binding_remove_vif(sc, in);
iwm_mvm_mac_ctxt_remove(sc, in);
return error;
#endif
}
static struct ieee80211_node *
iwm_node_alloc(struct ieee80211vap *vap, const uint8_t mac[IEEE80211_ADDR_LEN])
{
return malloc(sizeof (struct iwm_node), M_80211_NODE,
M_NOWAIT | M_ZERO);
}
static void
iwm_setrates(struct iwm_softc *sc, struct iwm_node *in)
{
struct ieee80211_node *ni = &in->in_ni;
struct iwm_lq_cmd *lq = &in->in_lq;
int nrates = ni->ni_rates.rs_nrates;
int i, ridx, tab = 0;
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
// int txant = 0;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (nrates > nitems(lq->rs_table)) {
device_printf(sc->sc_dev,
"%s: node supports %d rates, driver handles "
"only %zu\n", __func__, nrates, nitems(lq->rs_table));
return;
}
[iwm] fix up the rate control setup code to initialise rates in the order we want to use it. The rate table was being initialised in low->high, but the link quality table was being initialised high->low. So, when we did a lookup, we would get the indexes wrong. This started by a patch from dragonflybsd which reversed how the ni->in_ridx[] array is being used; I'd rather it all be consistent. So, this is consistent. Inspired by: what I did to iwn(4) a while ago Inspired by: DragonflyBSD; <imre@vdsz.com>
2016-01-18 05:43:34 +00:00
if (nrates == 0) {
device_printf(sc->sc_dev,
"%s: node supports 0 rates, odd!\n", __func__);
return;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* XXX .. and most of iwm_node is not initialised explicitly;
* it's all just 0x0 passed to the firmware.
*/
/* first figure out which rates we should support */
/* XXX TODO: this isn't 11n aware /at all/ */
memset(&in->in_ridx, -1, sizeof(in->in_ridx));
IWM_DPRINTF(sc, IWM_DEBUG_TXRATE,
"%s: nrates=%d\n", __func__, nrates);
[iwm] fix up the rate control setup code to initialise rates in the order we want to use it. The rate table was being initialised in low->high, but the link quality table was being initialised high->low. So, when we did a lookup, we would get the indexes wrong. This started by a patch from dragonflybsd which reversed how the ni->in_ridx[] array is being used; I'd rather it all be consistent. So, this is consistent. Inspired by: what I did to iwn(4) a while ago Inspired by: DragonflyBSD; <imre@vdsz.com>
2016-01-18 05:43:34 +00:00
/*
* Loop over nrates and populate in_ridx from the highest
* rate to the lowest rate. Remember, in_ridx[] has
* IEEE80211_RATE_MAXSIZE entries!
*/
for (i = 0; i < min(nrates, IEEE80211_RATE_MAXSIZE); i++) {
int rate = ni->ni_rates.rs_rates[(nrates - 1) - i] & IEEE80211_RATE_VAL;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Map 802.11 rate to HW rate index. */
for (ridx = 0; ridx <= IWM_RIDX_MAX; ridx++)
if (iwm_rates[ridx].rate == rate)
break;
if (ridx > IWM_RIDX_MAX) {
device_printf(sc->sc_dev,
"%s: WARNING: device rate for %d not found!\n",
__func__, rate);
} else {
IWM_DPRINTF(sc, IWM_DEBUG_TXRATE,
"%s: rate: i: %d, rate=%d, ridx=%d\n",
__func__,
i,
rate,
ridx);
in->in_ridx[i] = ridx;
}
}
/* then construct a lq_cmd based on those */
memset(lq, 0, sizeof(*lq));
lq->sta_id = IWM_STATION_ID;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/* For HT, always enable RTS/CTS to avoid excessive retries. */
if (ni->ni_flags & IEEE80211_NODE_HT)
lq->flags |= IWM_LQ_FLAG_USE_RTS_MSK;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* are these used? (we don't do SISO or MIMO)
* need to set them to non-zero, though, or we get an error.
*/
lq->single_stream_ant_msk = 1;
lq->dual_stream_ant_msk = 1;
/*
* Build the actual rate selection table.
* The lowest bits are the rates. Additionally,
* CCK needs bit 9 to be set. The rest of the bits
* we add to the table select the tx antenna
* Note that we add the rates in the highest rate first
* (opposite of ni_rates).
*/
/*
* XXX TODO: this should be looping over the min of nrates
* and LQ_MAX_RETRY_NUM. Sigh.
*/
for (i = 0; i < nrates; i++) {
int nextant;
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
#if 0
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (txant == 0)
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
txant = iwm_fw_valid_tx_ant(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
nextant = 1<<(ffs(txant)-1);
txant &= ~nextant;
[iwm] fix default antenna selection a bit; fix 5GHz rate control handling. * Don't do the antenna switching when setting up the rate table - we don't take into account whether it's an active antenna or not (eg shared with BT.) I'll look into this a bit more later. * The default antenna is still 1, I'll look into that a bit more later. (So no, this doesn't fix it for Larry who needs ANT_B to be active, not ANT_A.) * My changes to the rate control setup used the wrong method for finding a suitable rate, which led to 1M CCK frames being queued for 11a operation. This is .. sub-optimal. Change the rate control lookup path to use the global table instead of the per-node table, which won't be setup until the node rate table is setup (which happens way too late in this game.) Tested: * Intel 7260, 2G and 5G operation.
2016-09-24 01:21:42 +00:00
#else
nextant = iwm_fw_valid_tx_ant(sc);
#endif
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* Map the rate id into a rate index into
* our hardware table containing the
* configuration to use for this rate.
*/
[iwm] fix up the rate control setup code to initialise rates in the order we want to use it. The rate table was being initialised in low->high, but the link quality table was being initialised high->low. So, when we did a lookup, we would get the indexes wrong. This started by a patch from dragonflybsd which reversed how the ni->in_ridx[] array is being used; I'd rather it all be consistent. So, this is consistent. Inspired by: what I did to iwn(4) a while ago Inspired by: DragonflyBSD; <imre@vdsz.com>
2016-01-18 05:43:34 +00:00
ridx = in->in_ridx[i];
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
tab = iwm_rates[ridx].plcp;
tab |= nextant << IWM_RATE_MCS_ANT_POS;
if (IWM_RIDX_IS_CCK(ridx))
tab |= IWM_RATE_MCS_CCK_MSK;
IWM_DPRINTF(sc, IWM_DEBUG_TXRATE,
"station rate i=%d, rate=%d, hw=%x\n",
i, iwm_rates[ridx].rate, tab);
lq->rs_table[i] = htole32(tab);
}
/* then fill the rest with the lowest possible rate */
for (i = nrates; i < nitems(lq->rs_table); i++) {
KASSERT(tab != 0, ("invalid tab"));
lq->rs_table[i] = htole32(tab);
}
}
static int
iwm_media_change(struct ifnet *ifp)
{
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct ieee80211vap *vap = ifp->if_softc;
struct ieee80211com *ic = vap->iv_ic;
struct iwm_softc *sc = ic->ic_softc;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
int error;
error = ieee80211_media_change(ifp);
if (error != ENETRESET)
return error;
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
IWM_LOCK(sc);
if (ic->ic_nrunning > 0) {
iwm_stop(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_init(sc);
}
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
IWM_UNLOCK(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
}
static int
iwm_newstate(struct ieee80211vap *vap, enum ieee80211_state nstate, int arg)
{
struct iwm_vap *ivp = IWM_VAP(vap);
struct ieee80211com *ic = vap->iv_ic;
Convert more ifp->if_softc -> ic->ic_softc. These should be a big no-op. Tested: * make universe
2015-08-17 23:44:38 +00:00
struct iwm_softc *sc = ic->ic_softc;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
struct iwm_node *in;
int error;
IWM_DPRINTF(sc, IWM_DEBUG_STATE,
"switching state %s -> %s\n",
ieee80211_state_name[vap->iv_state],
ieee80211_state_name[nstate]);
IEEE80211_UNLOCK(ic);
IWM_LOCK(sc);
[iwm] add LED blinking for iwm hardware that supports it. Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD, Linux iwlwifi/mvm
2016-06-02 04:42:28 +00:00
if (vap->iv_state == IEEE80211_S_SCAN && nstate != vap->iv_state)
iwm_led_blink_stop(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* disable beacon filtering if we're hopping out of RUN */
if (vap->iv_state == IEEE80211_S_RUN && nstate != vap->iv_state) {
iwm_mvm_disable_beacon_filter(sc);
iwm: reorganize if_iwmvar.h - Change order of data in if_iwmvar.h (like it is in other drivers: defines, data structures, vap/node structures, softc struct and locks); use indentation. - Fix IWM_LOCK(_sc) / IWM_UNLOCK(_sc) macro. - Add IWM_LOCK_INIT / DESTROY(sc) + fix mtx_init() usage. - Wrap iwm_node casts into IWM_NODE() macro. - Drop some fields: * wt_hwqueue from Tx radiotap header; * macaddr[6] from iwm_vap; Approved by: adrian Differential Revision: https://reviews.freebsd.org/D4753
2016-01-03 10:06:10 +00:00
if (((in = IWM_NODE(vap->iv_bss)) != NULL))
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
in->in_assoc = 0;
[iwm] Improve reliability of iwm_release on disassociation a bit. * We need to first call ivp->iv_newstate(), to enqueue the deauth/deassoc mgmt frame, then flush the tx queue, before actually calling iwm_release(). * cycling a wlan connection via wpa_cli frontend to wpa_supplicant, by issuing disconnect and reconnect commandos works quite well. (There is still an issue when disconnecting/reconnecting too quickly) * Reassociating or roaming via wpa_supplicant is still broken. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7943
2016-09-19 22:36:30 +00:00
if (nstate == IEEE80211_S_INIT) {
IWM_UNLOCK(sc);
IEEE80211_LOCK(ic);
error = ivp->iv_newstate(vap, nstate, arg);
IEEE80211_UNLOCK(ic);
IWM_LOCK(sc);
iwm_release(sc, NULL);
IWM_UNLOCK(sc);
IEEE80211_LOCK(ic);
return error;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* It's impossible to directly go RUN->SCAN. If we iwm_release()
* above then the card will be completely reinitialized,
* so the driver must do everything necessary to bring the card
* from INIT to SCAN.
*
* Additionally, upon receiving deauth frame from AP,
* OpenBSD 802.11 stack puts the driver in IEEE80211_S_AUTH
* state. This will also fail with this driver, so bring the FSM
* from IEEE80211_S_RUN to IEEE80211_S_SCAN in this case as well.
*
* XXX TODO: fix this for FreeBSD!
*/
if (nstate == IEEE80211_S_SCAN ||
nstate == IEEE80211_S_AUTH ||
nstate == IEEE80211_S_ASSOC) {
IWM_DPRINTF(sc, IWM_DEBUG_STATE,
"Force transition to INIT; MGT=%d\n", arg);
IWM_UNLOCK(sc);
IEEE80211_LOCK(ic);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/* Always pass arg as -1 since we can't Tx right now. */
/*
* XXX arg is just ignored anyway when transitioning
* to IEEE80211_S_INIT.
*/
vap->iv_newstate(vap, IEEE80211_S_INIT, -1);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_STATE,
"Going INIT->SCAN\n");
nstate = IEEE80211_S_SCAN;
IEEE80211_UNLOCK(ic);
IWM_LOCK(sc);
}
}
switch (nstate) {
case IEEE80211_S_INIT:
break;
case IEEE80211_S_AUTH:
if ((error = iwm_auth(vap, sc)) != 0) {
device_printf(sc->sc_dev,
"%s: could not move to auth state: %d\n",
__func__, error);
break;
}
break;
case IEEE80211_S_ASSOC:
if ((error = iwm_assoc(vap, sc)) != 0) {
device_printf(sc->sc_dev,
"%s: failed to associate: %d\n", __func__,
error);
break;
}
break;
case IEEE80211_S_RUN:
{
struct iwm_host_cmd cmd = {
.id = IWM_LQ_CMD,
.len = { sizeof(in->in_lq), },
.flags = IWM_CMD_SYNC,
};
/* Update the association state, now we have it all */
/* (eg associd comes in at this point */
error = iwm_assoc(vap, sc);
if (error != 0) {
device_printf(sc->sc_dev,
"%s: failed to update association state: %d\n",
__func__,
error);
break;
}
iwm: reorganize if_iwmvar.h - Change order of data in if_iwmvar.h (like it is in other drivers: defines, data structures, vap/node structures, softc struct and locks); use indentation. - Fix IWM_LOCK(_sc) / IWM_UNLOCK(_sc) macro. - Add IWM_LOCK_INIT / DESTROY(sc) + fix mtx_init() usage. - Wrap iwm_node casts into IWM_NODE() macro. - Drop some fields: * wt_hwqueue from Tx radiotap header; * macaddr[6] from iwm_vap; Approved by: adrian Differential Revision: https://reviews.freebsd.org/D4753
2016-01-03 10:06:10 +00:00
in = IWM_NODE(vap->iv_bss);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_mvm_power_mac_update_mode(sc, in);
iwm_mvm_enable_beacon_filter(sc, in);
iwm_mvm_update_quotas(sc, in);
iwm_setrates(sc, in);
cmd.data[0] = &in->in_lq;
if ((error = iwm_send_cmd(sc, &cmd)) != 0) {
device_printf(sc->sc_dev,
"%s: IWM_LQ_CMD failed\n", __func__);
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
iwm_mvm_led_enable(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
break;
}
default:
break;
}
IWM_UNLOCK(sc);
IEEE80211_LOCK(ic);
return (ivp->iv_newstate(vap, nstate, arg));
}
void
iwm_endscan_cb(void *arg, int pending)
{
struct iwm_softc *sc = arg;
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct ieee80211com *ic = &sc->sc_ic;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_SCAN | IWM_DEBUG_TRACE,
"%s: scan ended\n",
__func__);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
ieee80211_scan_done(TAILQ_FIRST(&ic->ic_vaps));
}
/*
* Aging and idle timeouts for the different possible scenarios
* in default configuration
*/
static const uint32_t
iwm_sf_full_timeout_def[IWM_SF_NUM_SCENARIO][IWM_SF_NUM_TIMEOUT_TYPES] = {
{
htole32(IWM_SF_SINGLE_UNICAST_AGING_TIMER_DEF),
htole32(IWM_SF_SINGLE_UNICAST_IDLE_TIMER_DEF)
},
{
htole32(IWM_SF_AGG_UNICAST_AGING_TIMER_DEF),
htole32(IWM_SF_AGG_UNICAST_IDLE_TIMER_DEF)
},
{
htole32(IWM_SF_MCAST_AGING_TIMER_DEF),
htole32(IWM_SF_MCAST_IDLE_TIMER_DEF)
},
{
htole32(IWM_SF_BA_AGING_TIMER_DEF),
htole32(IWM_SF_BA_IDLE_TIMER_DEF)
},
{
htole32(IWM_SF_TX_RE_AGING_TIMER_DEF),
htole32(IWM_SF_TX_RE_IDLE_TIMER_DEF)
},
};
/*
* Aging and idle timeouts for the different possible scenarios
* in single BSS MAC configuration.
*/
static const uint32_t
iwm_sf_full_timeout[IWM_SF_NUM_SCENARIO][IWM_SF_NUM_TIMEOUT_TYPES] = {
{
htole32(IWM_SF_SINGLE_UNICAST_AGING_TIMER),
htole32(IWM_SF_SINGLE_UNICAST_IDLE_TIMER)
},
{
htole32(IWM_SF_AGG_UNICAST_AGING_TIMER),
htole32(IWM_SF_AGG_UNICAST_IDLE_TIMER)
},
{
htole32(IWM_SF_MCAST_AGING_TIMER),
htole32(IWM_SF_MCAST_IDLE_TIMER)
},
{
htole32(IWM_SF_BA_AGING_TIMER),
htole32(IWM_SF_BA_IDLE_TIMER)
},
{
htole32(IWM_SF_TX_RE_AGING_TIMER),
htole32(IWM_SF_TX_RE_IDLE_TIMER)
},
};
static void
iwm_mvm_fill_sf_command(struct iwm_softc *sc, struct iwm_sf_cfg_cmd *sf_cmd,
struct ieee80211_node *ni)
{
int i, j, watermark;
sf_cmd->watermark[IWM_SF_LONG_DELAY_ON] = htole32(IWM_SF_W_MARK_SCAN);
/*
* If we are in association flow - check antenna configuration
* capabilities of the AP station, and choose the watermark accordingly.
*/
if (ni) {
if (ni->ni_flags & IEEE80211_NODE_HT) {
#ifdef notyet
if (ni->ni_rxmcs[2] != 0)
watermark = IWM_SF_W_MARK_MIMO3;
else if (ni->ni_rxmcs[1] != 0)
watermark = IWM_SF_W_MARK_MIMO2;
else
#endif
watermark = IWM_SF_W_MARK_SISO;
} else {
watermark = IWM_SF_W_MARK_LEGACY;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/* default watermark value for unassociated mode. */
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
} else {
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
watermark = IWM_SF_W_MARK_MIMO2;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sf_cmd->watermark[IWM_SF_FULL_ON] = htole32(watermark);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
for (i = 0; i < IWM_SF_NUM_SCENARIO; i++) {
for (j = 0; j < IWM_SF_NUM_TIMEOUT_TYPES; j++) {
sf_cmd->long_delay_timeouts[i][j] =
htole32(IWM_SF_LONG_DELAY_AGING_TIMER);
}
}
if (ni) {
memcpy(sf_cmd->full_on_timeouts, iwm_sf_full_timeout,
sizeof(iwm_sf_full_timeout));
} else {
memcpy(sf_cmd->full_on_timeouts, iwm_sf_full_timeout_def,
sizeof(iwm_sf_full_timeout_def));
}
}
static int
iwm_mvm_sf_config(struct iwm_softc *sc, enum iwm_sf_state new_state)
{
struct ieee80211com *ic = &sc->sc_ic;
struct ieee80211vap *vap = TAILQ_FIRST(&ic->ic_vaps);
struct iwm_sf_cfg_cmd sf_cmd = {
.state = htole32(IWM_SF_FULL_ON),
};
int ret = 0;
if (sc->sc_device_family == IWM_DEVICE_FAMILY_8000)
sf_cmd.state |= htole32(IWM_SF_CFG_DUMMY_NOTIF_OFF);
switch (new_state) {
case IWM_SF_UNINIT:
case IWM_SF_INIT_OFF:
iwm_mvm_fill_sf_command(sc, &sf_cmd, NULL);
break;
case IWM_SF_FULL_ON:
iwm_mvm_fill_sf_command(sc, &sf_cmd, vap->iv_bss);
break;
default:
IWM_DPRINTF(sc, IWM_DEBUG_PWRSAVE,
"Invalid state: %d. not sending Smart Fifo cmd\n",
new_state);
return EINVAL;
}
ret = iwm_mvm_send_cmd_pdu(sc, IWM_REPLY_SF_CFG_CMD, IWM_CMD_ASYNC,
sizeof(sf_cmd), &sf_cmd);
return ret;
}
static int
iwm_send_bt_init_conf(struct iwm_softc *sc)
{
struct iwm_bt_coex_cmd bt_cmd;
bt_cmd.mode = htole32(IWM_BT_COEX_WIFI);
bt_cmd.enabled_modules = htole32(IWM_BT_COEX_HIGH_BAND_RET);
return iwm_mvm_send_cmd_pdu(sc, IWM_BT_CONFIG, 0, sizeof(bt_cmd),
&bt_cmd);
}
static int
iwm_send_update_mcc_cmd(struct iwm_softc *sc, const char *alpha2)
{
struct iwm_mcc_update_cmd mcc_cmd;
struct iwm_host_cmd hcmd = {
.id = IWM_MCC_UPDATE_CMD,
.flags = (IWM_CMD_SYNC | IWM_CMD_WANT_SKB),
.data = { &mcc_cmd },
};
int ret;
#ifdef IWM_DEBUG
struct iwm_rx_packet *pkt;
struct iwm_mcc_update_resp_v1 *mcc_resp_v1 = NULL;
struct iwm_mcc_update_resp *mcc_resp;
int n_channels;
uint16_t mcc;
#endif
int resp_v2 = isset(sc->sc_enabled_capa,
IWM_UCODE_TLV_CAPA_LAR_SUPPORT_V2);
memset(&mcc_cmd, 0, sizeof(mcc_cmd));
mcc_cmd.mcc = htole16(alpha2[0] << 8 | alpha2[1]);
if ((sc->sc_ucode_api & IWM_UCODE_TLV_API_WIFI_MCC_UPDATE) ||
isset(sc->sc_enabled_capa, IWM_UCODE_TLV_CAPA_LAR_MULTI_MCC))
mcc_cmd.source_id = IWM_MCC_SOURCE_GET_CURRENT;
else
mcc_cmd.source_id = IWM_MCC_SOURCE_OLD_FW;
if (resp_v2)
hcmd.len[0] = sizeof(struct iwm_mcc_update_cmd);
else
hcmd.len[0] = sizeof(struct iwm_mcc_update_cmd_v1);
IWM_DPRINTF(sc, IWM_DEBUG_NODE,
"send MCC update to FW with '%c%c' src = %d\n",
alpha2[0], alpha2[1], mcc_cmd.source_id);
ret = iwm_send_cmd(sc, &hcmd);
if (ret)
return ret;
#ifdef IWM_DEBUG
pkt = hcmd.resp_pkt;
/* Extract MCC response */
if (resp_v2) {
mcc_resp = (void *)pkt->data;
mcc = mcc_resp->mcc;
n_channels = le32toh(mcc_resp->n_channels);
} else {
mcc_resp_v1 = (void *)pkt->data;
mcc = mcc_resp_v1->mcc;
n_channels = le32toh(mcc_resp_v1->n_channels);
}
/* W/A for a FW/NVM issue - returns 0x00 for the world domain */
if (mcc == 0)
mcc = 0x3030; /* "00" - world */
IWM_DPRINTF(sc, IWM_DEBUG_NODE,
"regulatory domain '%c%c' (%d channels available)\n",
mcc >> 8, mcc & 0xff, n_channels);
#endif
iwm_free_resp(sc, &hcmd);
return 0;
}
static void
iwm_mvm_tt_tx_backoff(struct iwm_softc *sc, uint32_t backoff)
{
struct iwm_host_cmd cmd = {
.id = IWM_REPLY_THERMAL_MNG_BACKOFF,
.len = { sizeof(uint32_t), },
.data = { &backoff, },
};
if (iwm_send_cmd(sc, &cmd) != 0) {
device_printf(sc->sc_dev,
"failed to change thermal tx backoff\n");
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
}
static int
iwm_init_hw(struct iwm_softc *sc)
{
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct ieee80211com *ic = &sc->sc_ic;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
int error, i, ac;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if ((error = iwm_start_hw(sc)) != 0) {
printf("iwm_start_hw: failed %d\n", error);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if ((error = iwm_run_init_mvm_ucode(sc, 0)) != 0) {
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
printf("iwm_run_init_mvm_ucode: failed %d\n", error);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return error;
}
/*
* should stop and start HW since that INIT
* image just loaded
*/
iwm_stop_device(sc);
if ((error = iwm_start_hw(sc)) != 0) {
device_printf(sc->sc_dev, "could not initialize hardware\n");
return error;
}
/* omstart, this time with the regular firmware */
error = iwm_mvm_load_ucode_wait_alive(sc, IWM_UCODE_TYPE_REGULAR);
if (error) {
device_printf(sc->sc_dev, "could not load firmware\n");
goto error;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if ((error = iwm_send_bt_init_conf(sc)) != 0) {
device_printf(sc->sc_dev, "bt init conf failed\n");
goto error;
}
if ((error = iwm_send_tx_ant_cfg(sc, iwm_fw_valid_tx_ant(sc))) != 0) {
device_printf(sc->sc_dev, "antenna config failed\n");
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
goto error;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Send phy db control command and then phy db calibration*/
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if ((error = iwm_send_phy_db_data(sc)) != 0) {
device_printf(sc->sc_dev, "phy_db_data failed\n");
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
goto error;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if ((error = iwm_send_phy_cfg_cmd(sc)) != 0) {
device_printf(sc->sc_dev, "phy_cfg_cmd failed\n");
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
goto error;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Add auxiliary station for scanning */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if ((error = iwm_mvm_add_aux_sta(sc)) != 0) {
device_printf(sc->sc_dev, "add_aux_sta failed\n");
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
goto error;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
for (i = 0; i < IWM_NUM_PHY_CTX; i++) {
/*
* The channel used here isn't relevant as it's
* going to be overwritten in the other flows.
* For now use the first channel we have.
*/
if ((error = iwm_mvm_phy_ctxt_add(sc,
&sc->sc_phyctxt[i], &ic->ic_channels[1], 1, 1)) != 0)
goto error;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/* Initialize tx backoffs to the minimum. */
if (sc->sc_device_family == IWM_DEVICE_FAMILY_7000)
iwm_mvm_tt_tx_backoff(sc, 0);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
error = iwm_mvm_power_update_device(sc);
if (error)
goto error;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (isset(sc->sc_enabled_capa, IWM_UCODE_TLV_CAPA_LAR_SUPPORT)) {
if ((error = iwm_send_update_mcc_cmd(sc, "ZZ")) != 0)
goto error;
}
if (isset(sc->sc_enabled_capa, IWM_UCODE_TLV_CAPA_UMAC_SCAN)) {
if ((error = iwm_mvm_config_umac_scan(sc)) != 0)
goto error;
}
/* Enable Tx queues. */
for (ac = 0; ac < WME_NUM_AC; ac++) {
error = iwm_enable_txq(sc, IWM_STATION_ID, ac,
iwm_mvm_ac_to_tx_fifo[ac]);
if (error)
goto error;
}
if ((error = iwm_mvm_disable_beacon_filter(sc)) != 0) {
device_printf(sc->sc_dev, "failed to disable beacon filter\n");
goto error;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
return 0;
error:
iwm_stop_device(sc);
return error;
}
/* Allow multicast from our BSSID. */
static int
iwm_allow_mcast(struct ieee80211vap *vap, struct iwm_softc *sc)
{
struct ieee80211_node *ni = vap->iv_bss;
struct iwm_mcast_filter_cmd *cmd;
size_t size;
int error;
size = roundup(sizeof(*cmd), 4);
cmd = malloc(size, M_DEVBUF, M_NOWAIT | M_ZERO);
if (cmd == NULL)
return ENOMEM;
cmd->filter_own = 1;
cmd->port_id = 0;
cmd->count = 0;
cmd->pass_all = 1;
IEEE80211_ADDR_COPY(cmd->bssid, ni->ni_bssid);
error = iwm_mvm_send_cmd_pdu(sc, IWM_MCAST_FILTER_CMD,
IWM_CMD_SYNC, size, cmd);
free(cmd, M_DEVBUF);
return (error);
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/*
* ifnet interfaces
*/
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static void
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
iwm_init(struct iwm_softc *sc)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
int error;
if (sc->sc_flags & IWM_FLAG_HW_INITED) {
return;
}
sc->sc_generation++;
sc->sc_flags &= ~IWM_FLAG_STOPPED;
if ((error = iwm_init_hw(sc)) != 0) {
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
printf("iwm_init_hw failed %d\n", error);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
iwm_stop(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return;
}
/*
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
* Ok, firmware loaded and we are jogging
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
*/
sc->sc_flags |= IWM_FLAG_HW_INITED;
callout_reset(&sc->sc_watchdog_to, hz, iwm_watchdog, sc);
}
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
static int
iwm_transmit(struct ieee80211com *ic, struct mbuf *m)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct iwm_softc *sc;
int error;
sc = ic->ic_softc;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_LOCK(sc);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
if ((sc->sc_flags & IWM_FLAG_HW_INITED) == 0) {
IWM_UNLOCK(sc);
return (ENXIO);
}
error = mbufq_enqueue(&sc->sc_snd, m);
if (error) {
IWM_UNLOCK(sc);
return (error);
}
iwm_start(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_UNLOCK(sc);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
return (0);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
/*
* Dequeue packets from sendq and call send.
*/
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static void
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
iwm_start(struct iwm_softc *sc)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
struct ieee80211_node *ni;
struct mbuf *m;
int ac = 0;
IWM_DPRINTF(sc, IWM_DEBUG_XMIT | IWM_DEBUG_TRACE, "->%s\n", __func__);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
while (sc->qfullmsk == 0 &&
(m = mbufq_dequeue(&sc->sc_snd)) != NULL) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ni = (struct ieee80211_node *)m->m_pkthdr.rcvif;
if (iwm_tx(sc, m, ni, ac) != 0) {
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
if_inc_counter(ni->ni_vap->iv_ifp,
IFCOUNTER_OERRORS, 1);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ieee80211_free_node(ni);
continue;
}
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
sc->sc_tx_timer = 15;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
IWM_DPRINTF(sc, IWM_DEBUG_XMIT | IWM_DEBUG_TRACE, "<-%s\n", __func__);
}
static void
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
iwm_stop(struct iwm_softc *sc)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
sc->sc_flags &= ~IWM_FLAG_HW_INITED;
sc->sc_flags |= IWM_FLAG_STOPPED;
sc->sc_generation++;
[iwm] add LED blinking for iwm hardware that supports it. Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD, Linux iwlwifi/mvm
2016-06-02 04:42:28 +00:00
iwm_led_blink_stop(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
sc->sc_tx_timer = 0;
iwm_stop_device(sc);
}
static void
iwm_watchdog(void *arg)
{
struct iwm_softc *sc = arg;
iwm: restart device after watchdog timeouts.
2016-05-19 22:43:21 +00:00
struct ieee80211com *ic = &sc->sc_ic;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (sc->sc_tx_timer > 0) {
if (--sc->sc_tx_timer == 0) {
device_printf(sc->sc_dev, "device timeout\n");
#ifdef IWM_DEBUG
iwm_nic_error(sc);
#endif
iwm: restart device after watchdog timeouts.
2016-05-19 22:43:21 +00:00
ieee80211_restart_all(ic);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
counter_u64_add(sc->sc_ic.ic_oerrors, 1);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return;
}
}
callout_reset(&sc->sc_watchdog_to, hz, iwm_watchdog, sc);
}
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
static void
iwm_parent(struct ieee80211com *ic)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct iwm_softc *sc = ic->ic_softc;
int startall = 0;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
IWM_LOCK(sc);
if (ic->ic_nrunning > 0) {
if (!(sc->sc_flags & IWM_FLAG_HW_INITED)) {
iwm_init(sc);
startall = 1;
}
} else if (sc->sc_flags & IWM_FLAG_HW_INITED)
iwm_stop(sc);
IWM_UNLOCK(sc);
if (startall)
ieee80211_start_all(ic);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
/*
* The interrupt side of things
*/
/*
* error dumping routines are from iwlwifi/mvm/utils.c
*/
/*
* Note: This structure is read from the device with IO accesses,
* and the reading already does the endian conversion. As it is
* read with uint32_t-sized accesses, any members with a different size
* need to be ordered correctly though!
*/
struct iwm_error_event_table {
uint32_t valid; /* (nonzero) valid, (0) log is empty */
uint32_t error_id; /* type of error */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
uint32_t trm_hw_status0; /* TRM HW status */
uint32_t trm_hw_status1; /* TRM HW status */
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
uint32_t blink2; /* branch link */
uint32_t ilink1; /* interrupt link */
uint32_t ilink2; /* interrupt link */
uint32_t data1; /* error-specific data */
uint32_t data2; /* error-specific data */
uint32_t data3; /* error-specific data */
uint32_t bcon_time; /* beacon timer */
uint32_t tsf_low; /* network timestamp function timer */
uint32_t tsf_hi; /* network timestamp function timer */
uint32_t gp1; /* GP1 timer register */
uint32_t gp2; /* GP2 timer register */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
uint32_t fw_rev_type; /* firmware revision type */
uint32_t major; /* uCode version major */
uint32_t minor; /* uCode version minor */
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
uint32_t hw_ver; /* HW Silicon version */
uint32_t brd_ver; /* HW board version */
uint32_t log_pc; /* log program counter */
uint32_t frame_ptr; /* frame pointer */
uint32_t stack_ptr; /* stack pointer */
uint32_t hcmd; /* last host command header */
uint32_t isr0; /* isr status register LMPM_NIC_ISR0:
* rxtx_flag */
uint32_t isr1; /* isr status register LMPM_NIC_ISR1:
* host_flag */
uint32_t isr2; /* isr status register LMPM_NIC_ISR2:
* enc_flag */
uint32_t isr3; /* isr status register LMPM_NIC_ISR3:
* time_flag */
uint32_t isr4; /* isr status register LMPM_NIC_ISR4:
* wico interrupt */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
uint32_t last_cmd_id; /* last HCMD id handled by the firmware */
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
uint32_t wait_event; /* wait event() caller address */
uint32_t l2p_control; /* L2pControlField */
uint32_t l2p_duration; /* L2pDurationField */
uint32_t l2p_mhvalid; /* L2pMhValidBits */
uint32_t l2p_addr_match; /* L2pAddrMatchStat */
uint32_t lmpm_pmg_sel; /* indicate which clocks are turned on
* (LMPM_PMG_SEL) */
uint32_t u_timestamp; /* indicate when the date and time of the
* compilation */
uint32_t flow_handler; /* FH read/write pointers, RX credit */
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
} __packed /* LOG_ERROR_TABLE_API_S_VER_3 */;
/*
* UMAC error struct - relevant starting from family 8000 chip.
* Note: This structure is read from the device with IO accesses,
* and the reading already does the endian conversion. As it is
* read with u32-sized accesses, any members with a different size
* need to be ordered correctly though!
*/
struct iwm_umac_error_event_table {
uint32_t valid; /* (nonzero) valid, (0) log is empty */
uint32_t error_id; /* type of error */
uint32_t blink1; /* branch link */
uint32_t blink2; /* branch link */
uint32_t ilink1; /* interrupt link */
uint32_t ilink2; /* interrupt link */
uint32_t data1; /* error-specific data */
uint32_t data2; /* error-specific data */
uint32_t data3; /* error-specific data */
uint32_t umac_major;
uint32_t umac_minor;
uint32_t frame_pointer; /* core register 27*/
uint32_t stack_pointer; /* core register 28 */
uint32_t cmd_header; /* latest host cmd sent to UMAC */
uint32_t nic_isr_pref; /* ISR status register */
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
} __packed;
#define ERROR_START_OFFSET (1 * sizeof(uint32_t))
#define ERROR_ELEM_SIZE (7 * sizeof(uint32_t))
#ifdef IWM_DEBUG
struct {
const char *name;
uint8_t num;
} advanced_lookup[] = {
{ "NMI_INTERRUPT_WDG", 0x34 },
{ "SYSASSERT", 0x35 },
{ "UCODE_VERSION_MISMATCH", 0x37 },
{ "BAD_COMMAND", 0x38 },
{ "NMI_INTERRUPT_DATA_ACTION_PT", 0x3C },
{ "FATAL_ERROR", 0x3D },
{ "NMI_TRM_HW_ERR", 0x46 },
{ "NMI_INTERRUPT_TRM", 0x4C },
{ "NMI_INTERRUPT_BREAK_POINT", 0x54 },
{ "NMI_INTERRUPT_WDG_RXF_FULL", 0x5C },
{ "NMI_INTERRUPT_WDG_NO_RBD_RXF_FULL", 0x64 },
{ "NMI_INTERRUPT_HOST", 0x66 },
{ "NMI_INTERRUPT_ACTION_PT", 0x7C },
{ "NMI_INTERRUPT_UNKNOWN", 0x84 },
{ "NMI_INTERRUPT_INST_ACTION_PT", 0x86 },
{ "ADVANCED_SYSASSERT", 0 },
};
static const char *
iwm_desc_lookup(uint32_t num)
{
int i;
for (i = 0; i < nitems(advanced_lookup) - 1; i++)
if (advanced_lookup[i].num == num)
return advanced_lookup[i].name;
/* No entry matches 'num', so it is the last: ADVANCED_SYSASSERT */
return advanced_lookup[i].name;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
static void
iwm_nic_umac_error(struct iwm_softc *sc)
{
struct iwm_umac_error_event_table table;
uint32_t base;
base = sc->sc_uc.uc_umac_error_event_table;
if (base < 0x800000) {
device_printf(sc->sc_dev, "Invalid error log pointer 0x%08x\n",
base);
return;
}
if (iwm_read_mem(sc, base, &table, sizeof(table)/sizeof(uint32_t))) {
device_printf(sc->sc_dev, "reading errlog failed\n");
return;
}
if (ERROR_START_OFFSET <= table.valid * ERROR_ELEM_SIZE) {
device_printf(sc->sc_dev, "Start UMAC Error Log Dump:\n");
device_printf(sc->sc_dev, "Status: 0x%x, count: %d\n",
sc->sc_flags, table.valid);
}
device_printf(sc->sc_dev, "0x%08X | %s\n", table.error_id,
iwm_desc_lookup(table.error_id));
device_printf(sc->sc_dev, "0x%08X | umac branchlink1\n", table.blink1);
device_printf(sc->sc_dev, "0x%08X | umac branchlink2\n", table.blink2);
device_printf(sc->sc_dev, "0x%08X | umac interruptlink1\n",
table.ilink1);
device_printf(sc->sc_dev, "0x%08X | umac interruptlink2\n",
table.ilink2);
device_printf(sc->sc_dev, "0x%08X | umac data1\n", table.data1);
device_printf(sc->sc_dev, "0x%08X | umac data2\n", table.data2);
device_printf(sc->sc_dev, "0x%08X | umac data3\n", table.data3);
device_printf(sc->sc_dev, "0x%08X | umac major\n", table.umac_major);
device_printf(sc->sc_dev, "0x%08X | umac minor\n", table.umac_minor);
device_printf(sc->sc_dev, "0x%08X | frame pointer\n",
table.frame_pointer);
device_printf(sc->sc_dev, "0x%08X | stack pointer\n",
table.stack_pointer);
device_printf(sc->sc_dev, "0x%08X | last host cmd\n", table.cmd_header);
device_printf(sc->sc_dev, "0x%08X | isr status reg\n",
table.nic_isr_pref);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* Support for dumping the error log seemed like a good idea ...
* but it's mostly hex junk and the only sensible thing is the
* hw/ucode revision (which we know anyway). Since it's here,
* I'll just leave it in, just in case e.g. the Intel guys want to
* help us decipher some "ADVANCED_SYSASSERT" later.
*/
static void
iwm_nic_error(struct iwm_softc *sc)
{
struct iwm_error_event_table table;
uint32_t base;
device_printf(sc->sc_dev, "dumping device error log\n");
base = sc->sc_uc.uc_error_event_table;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (base < 0x800000) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
"Invalid error log pointer 0x%08x\n", base);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (iwm_read_mem(sc, base, &table, sizeof(table)/sizeof(uint32_t))) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev, "reading errlog failed\n");
return;
}
if (!table.valid) {
device_printf(sc->sc_dev, "errlog not found, skipping\n");
return;
}
if (ERROR_START_OFFSET <= table.valid * ERROR_ELEM_SIZE) {
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
device_printf(sc->sc_dev, "Start Error Log Dump:\n");
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev, "Status: 0x%x, count: %d\n",
sc->sc_flags, table.valid);
}
device_printf(sc->sc_dev, "0x%08X | %-28s\n", table.error_id,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
iwm_desc_lookup(table.error_id));
device_printf(sc->sc_dev, "%08X | trm_hw_status0\n",
table.trm_hw_status0);
device_printf(sc->sc_dev, "%08X | trm_hw_status1\n",
table.trm_hw_status1);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev, "%08X | branchlink2\n", table.blink2);
device_printf(sc->sc_dev, "%08X | interruptlink1\n", table.ilink1);
device_printf(sc->sc_dev, "%08X | interruptlink2\n", table.ilink2);
device_printf(sc->sc_dev, "%08X | data1\n", table.data1);
device_printf(sc->sc_dev, "%08X | data2\n", table.data2);
device_printf(sc->sc_dev, "%08X | data3\n", table.data3);
device_printf(sc->sc_dev, "%08X | beacon time\n", table.bcon_time);
device_printf(sc->sc_dev, "%08X | tsf low\n", table.tsf_low);
device_printf(sc->sc_dev, "%08X | tsf hi\n", table.tsf_hi);
device_printf(sc->sc_dev, "%08X | time gp1\n", table.gp1);
device_printf(sc->sc_dev, "%08X | time gp2\n", table.gp2);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
device_printf(sc->sc_dev, "%08X | uCode revision type\n",
table.fw_rev_type);
device_printf(sc->sc_dev, "%08X | uCode version major\n", table.major);
device_printf(sc->sc_dev, "%08X | uCode version minor\n", table.minor);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev, "%08X | hw version\n", table.hw_ver);
device_printf(sc->sc_dev, "%08X | board version\n", table.brd_ver);
device_printf(sc->sc_dev, "%08X | hcmd\n", table.hcmd);
device_printf(sc->sc_dev, "%08X | isr0\n", table.isr0);
device_printf(sc->sc_dev, "%08X | isr1\n", table.isr1);
device_printf(sc->sc_dev, "%08X | isr2\n", table.isr2);
device_printf(sc->sc_dev, "%08X | isr3\n", table.isr3);
device_printf(sc->sc_dev, "%08X | isr4\n", table.isr4);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
device_printf(sc->sc_dev, "%08X | last cmd Id\n", table.last_cmd_id);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev, "%08X | wait_event\n", table.wait_event);
device_printf(sc->sc_dev, "%08X | l2p_control\n", table.l2p_control);
device_printf(sc->sc_dev, "%08X | l2p_duration\n", table.l2p_duration);
device_printf(sc->sc_dev, "%08X | l2p_mhvalid\n", table.l2p_mhvalid);
device_printf(sc->sc_dev, "%08X | l2p_addr_match\n", table.l2p_addr_match);
device_printf(sc->sc_dev, "%08X | lmpm_pmg_sel\n", table.lmpm_pmg_sel);
device_printf(sc->sc_dev, "%08X | timestamp\n", table.u_timestamp);
device_printf(sc->sc_dev, "%08X | flow_handler\n", table.flow_handler);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (sc->sc_uc.uc_umac_error_event_table)
iwm_nic_umac_error(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
#endif
#define ADVANCE_RXQ(sc) (sc->rxq.cur = (sc->rxq.cur + 1) % IWM_RX_RING_COUNT);
/*
* Process an IWM_CSR_INT_BIT_FH_RX or IWM_CSR_INT_BIT_SW_RX interrupt.
* Basic structure from if_iwn
*/
static void
iwm_notif_intr(struct iwm_softc *sc)
{
iwm(4): switch to ieee80211_runtask() Submitted by: Andiry Voskoboinyk (s3erios@gmail.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D5054
2016-07-25 23:44:44 +00:00
struct ieee80211com *ic = &sc->sc_ic;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
uint16_t hw;
bus_dmamap_sync(sc->rxq.stat_dma.tag, sc->rxq.stat_dma.map,
BUS_DMASYNC_POSTREAD);
hw = le16toh(sc->rxq.stat->closed_rb_num) & 0xfff;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/*
* Process responses
*/
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
while (sc->rxq.cur != hw) {
struct iwm_rx_ring *ring = &sc->rxq;
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
struct iwm_rx_data *data = &ring->data[ring->cur];
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
struct iwm_rx_packet *pkt;
struct iwm_cmd_response *cresp;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
int qid, idx, code;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
bus_dmamap_sync(ring->data_dmat, data->map,
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
BUS_DMASYNC_POSTREAD);
pkt = mtod(data->m, struct iwm_rx_packet *);
qid = pkt->hdr.qid & ~0x80;
idx = pkt->hdr.idx;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
code = IWM_WIDE_ID(pkt->hdr.flags, pkt->hdr.code);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_INTR,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
"rx packet qid=%d idx=%d type=%x %d %d\n",
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
pkt->hdr.qid & ~0x80, pkt->hdr.idx, code, ring->cur, hw);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* randomly get these from the firmware, no idea why.
* they at least seem harmless, so just ignore them for now
*/
if (__predict_false((pkt->hdr.code == 0 && qid == 0 && idx == 0)
|| pkt->len_n_flags == htole32(0x55550000))) {
ADVANCE_RXQ(sc);
continue;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
switch (code) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
case IWM_REPLY_RX_PHY_CMD:
iwm_mvm_rx_rx_phy_cmd(sc, pkt, data);
break;
case IWM_REPLY_RX_MPDU_CMD:
iwm_mvm_rx_rx_mpdu(sc, pkt, data);
break;
case IWM_TX_CMD:
iwm_mvm_rx_tx_cmd(sc, pkt, data);
break;
case IWM_MISSED_BEACONS_NOTIFICATION: {
struct iwm_missed_beacons_notif *resp;
int missed;
/* XXX look at mac_id to determine interface ID */
struct ieee80211vap *vap = TAILQ_FIRST(&ic->ic_vaps);
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
resp = (void *)pkt->data;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
missed = le32toh(resp->consec_missed_beacons);
IWM_DPRINTF(sc, IWM_DEBUG_BEACON | IWM_DEBUG_STATE,
"%s: MISSED_BEACON: mac_id=%d, "
"consec_since_last_rx=%d, consec=%d, num_expect=%d "
"num_rx=%d\n",
__func__,
le32toh(resp->mac_id),
le32toh(resp->consec_missed_beacons_since_last_rx),
le32toh(resp->consec_missed_beacons),
le32toh(resp->num_expected_beacons),
le32toh(resp->num_recvd_beacons));
/* Be paranoid */
if (vap == NULL)
break;
/* XXX no net80211 locking? */
if (vap->iv_state == IEEE80211_S_RUN &&
(ic->ic_flags & IEEE80211_F_SCAN) == 0) {
if (missed > vap->iv_bmissthreshold) {
/* XXX bad locking; turn into task */
IWM_UNLOCK(sc);
ieee80211_beacon_miss(ic);
IWM_LOCK(sc);
}
}
break; }
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
case IWM_MFUART_LOAD_NOTIFICATION:
break;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
case IWM_MVM_ALIVE: {
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
struct iwm_mvm_alive_resp_v1 *resp1;
struct iwm_mvm_alive_resp_v2 *resp2;
struct iwm_mvm_alive_resp_v3 *resp3;
if (iwm_rx_packet_payload_len(pkt) == sizeof(*resp1)) {
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
resp1 = (void *)pkt->data;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sc->sc_uc.uc_error_event_table
= le32toh(resp1->error_event_table_ptr);
sc->sc_uc.uc_log_event_table
= le32toh(resp1->log_event_table_ptr);
sc->sched_base = le32toh(resp1->scd_base_ptr);
if (resp1->status == IWM_ALIVE_STATUS_OK)
sc->sc_uc.uc_ok = 1;
else
sc->sc_uc.uc_ok = 0;
}
if (iwm_rx_packet_payload_len(pkt) == sizeof(*resp2)) {
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
resp2 = (void *)pkt->data;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sc->sc_uc.uc_error_event_table
= le32toh(resp2->error_event_table_ptr);
sc->sc_uc.uc_log_event_table
= le32toh(resp2->log_event_table_ptr);
sc->sched_base = le32toh(resp2->scd_base_ptr);
sc->sc_uc.uc_umac_error_event_table
= le32toh(resp2->error_info_addr);
if (resp2->status == IWM_ALIVE_STATUS_OK)
sc->sc_uc.uc_ok = 1;
else
sc->sc_uc.uc_ok = 0;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (iwm_rx_packet_payload_len(pkt) == sizeof(*resp3)) {
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
resp3 = (void *)pkt->data;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sc->sc_uc.uc_error_event_table
= le32toh(resp3->error_event_table_ptr);
sc->sc_uc.uc_log_event_table
= le32toh(resp3->log_event_table_ptr);
sc->sched_base = le32toh(resp3->scd_base_ptr);
sc->sc_uc.uc_umac_error_event_table
= le32toh(resp3->error_info_addr);
if (resp3->status == IWM_ALIVE_STATUS_OK)
sc->sc_uc.uc_ok = 1;
else
sc->sc_uc.uc_ok = 0;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
sc->sc_uc.uc_intr = 1;
wakeup(&sc->sc_uc);
break; }
case IWM_CALIB_RES_NOTIF_PHY_DB: {
struct iwm_calib_res_notif_phy_db *phy_db_notif;
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
phy_db_notif = (void *)pkt->data;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_phy_db_set_section(sc, phy_db_notif);
break; }
case IWM_STATISTICS_NOTIFICATION: {
struct iwm_notif_statistics *stats;
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
stats = (void *)pkt->data;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
memcpy(&sc->sc_stats, stats, sizeof(sc->sc_stats));
[iwm] fix up RSSI calculations for both scan results and normal RX operations. * hard code a noise floor of -96 for now. The noise floor update code returns some "interesting" values that I can't map to anything useful right now. * Ensure a default noise floor is set - otherwise the initial scan results have a noise floor of '0'. * Fix up the RSSI calculation to be correctly relative to the noise floor. The RSSI routines return an absolute value in dBm - so fix this up. * Cap RSSI values appropriately. * Ensure we pass in a 1/2 dB unit value in to net80211. Tested: * Intel 7260, STA mode iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: hw rev 0x140, fw ver 16.242414.0, address xx:xx:xx:xx:xx:xx
2016-09-18 05:07:18 +00:00
sc->sc_noise = iwm_get_noise(sc, &stats->rx.general);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
break; }
case IWM_NVM_ACCESS_CMD:
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
case IWM_MCC_UPDATE_CMD:
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (sc->sc_wantresp == ((qid << 16) | idx)) {
memcpy(sc->sc_cmd_resp,
pkt, sizeof(sc->sc_cmd_resp));
}
break;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
case IWM_MCC_CHUB_UPDATE_CMD: {
struct iwm_mcc_chub_notif *notif;
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
notif = (void *)pkt->data;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sc->sc_fw_mcc[0] = (notif->mcc & 0xff00) >> 8;
sc->sc_fw_mcc[1] = notif->mcc & 0xff;
sc->sc_fw_mcc[2] = '\0';
IWM_DPRINTF(sc, IWM_DEBUG_RESET,
"fw source %d sent CC '%s'\n",
notif->source_id, sc->sc_fw_mcc);
break; }
case IWM_DTS_MEASUREMENT_NOTIFICATION:
break;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
case IWM_PHY_CONFIGURATION_CMD:
case IWM_TX_ANT_CONFIGURATION_CMD:
case IWM_ADD_STA:
case IWM_MAC_CONTEXT_CMD:
case IWM_REPLY_SF_CFG_CMD:
case IWM_POWER_TABLE_CMD:
case IWM_PHY_CONTEXT_CMD:
case IWM_BINDING_CONTEXT_CMD:
case IWM_TIME_EVENT_CMD:
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
case IWM_WIDE_ID(IWM_ALWAYS_LONG_GROUP, IWM_SCAN_CFG_CMD):
case IWM_WIDE_ID(IWM_ALWAYS_LONG_GROUP, IWM_SCAN_REQ_UMAC):
case IWM_SCAN_OFFLOAD_REQUEST_CMD:
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
case IWM_REPLY_BEACON_FILTERING_CMD:
case IWM_MAC_PM_POWER_TABLE:
case IWM_TIME_QUOTA_CMD:
case IWM_REMOVE_STA:
case IWM_TXPATH_FLUSH:
case IWM_LQ_CMD:
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
case IWM_BT_CONFIG:
case IWM_REPLY_THERMAL_MNG_BACKOFF:
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
cresp = (void *)pkt->data;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (sc->sc_wantresp == ((qid << 16) | idx)) {
memcpy(sc->sc_cmd_resp,
pkt, sizeof(*pkt)+sizeof(*cresp));
}
break;
/* ignore */
case 0x6c: /* IWM_PHY_DB_CMD, no idea why it's not in fw-api.h */
break;
case IWM_INIT_COMPLETE_NOTIF:
sc->sc_init_complete = 1;
wakeup(&sc->sc_init_complete);
break;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
case IWM_SCAN_OFFLOAD_COMPLETE: {
struct iwm_periodic_scan_complete *notif;
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
notif = (void *)pkt->data;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
break;
}
case IWM_SCAN_ITERATION_COMPLETE: {
struct iwm_lmac_scan_complete_notif *notif;
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
notif = (void *)pkt->data;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
ieee80211_runtask(&sc->sc_ic, &sc->sc_es_task);
break;
}
case IWM_SCAN_COMPLETE_UMAC: {
struct iwm_umac_scan_complete *notif;
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
notif = (void *)pkt->data;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_SCAN,
"UMAC scan complete, status=0x%x\n",
notif->status);
#if 0 /* XXX This would be a duplicate scan end call */
taskqueue_enqueue(sc->sc_tq, &sc->sc_es_task);
#endif
break;
}
case IWM_SCAN_ITERATION_COMPLETE_UMAC: {
struct iwm_umac_scan_iter_complete_notif *notif;
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
notif = (void *)pkt->data;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_SCAN, "UMAC scan iteration "
"complete, status=0x%x, %d channels scanned\n",
notif->status, notif->scanned_channels);
ieee80211_runtask(&sc->sc_ic, &sc->sc_es_task);
break;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
case IWM_REPLY_ERROR: {
struct iwm_error_resp *resp;
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
resp = (void *)pkt->data;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev,
"firmware error 0x%x, cmd 0x%x\n",
le32toh(resp->error_type),
resp->cmd_id);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
break;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
case IWM_TIME_EVENT_NOTIFICATION: {
struct iwm_time_event_notif *notif;
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
notif = (void *)pkt->data;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_INTR,
[iwm] remove dead code. if_iwm - GC some dead code, left by a partially applied OpenBSD change. Taken-From: OpenBSD (if_iwm.c r1.69) Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD git 07dfed32ea39b980b0b80d27ff938e7c3ca4c0b5
2016-05-27 07:10:11 +00:00
"TE notif status = 0x%x action = 0x%x\n",
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
notif->status, notif->action);
break;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
case IWM_MCAST_FILTER_CMD:
break;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
case IWM_SCD_QUEUE_CFG: {
struct iwm_scd_txq_cfg_rsp *rsp;
[iwm] Get rid of SYNC_RESP_STRUCT and SYNC_RESP_PTR macros. * SYNC_RESP_STRUCT and SYNC_RESP_PTR originate from the OpenBSD version of iwm, and they weren't serving any real purpose in the FreeBSD port. * We just do a single bus_dmamap_sync for syncing the complete received frame, instead of explicitly bus_dmamap_sync-ing subranges of the frame like in the OpenBSD iwm code. Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7939
2016-11-01 22:38:26 +00:00
rsp = (void *)pkt->data;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_CMD,
"queue cfg token=0x%x sta_id=%d "
"tid=%d scd_queue=%d\n",
rsp->token, rsp->sta_id, rsp->tid,
rsp->scd_queue);
break;
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
default:
device_printf(sc->sc_dev,
"frame %d/%d %x UNHANDLED (this should "
"not happen)\n", qid, idx,
pkt->len_n_flags);
break;
}
/*
* Why test bit 0x80? The Linux driver:
*
* There is one exception: uCode sets bit 15 when it
* originates the response/notification, i.e. when the
* response/notification is not a direct response to a
* command sent by the driver. For example, uCode issues
* IWM_REPLY_RX when it sends a received frame to the driver;
* it is not a direct response to any driver command.
*
* Ok, so since when is 7 == 15? Well, the Linux driver
* uses a slightly different format for pkt->hdr, and "qid"
* is actually the upper byte of a two-byte field.
*/
if (!(pkt->hdr.qid & (1 << 7))) {
iwm_cmd_done(sc, pkt);
}
ADVANCE_RXQ(sc);
}
IWM_CLRBITS(sc, IWM_CSR_GP_CNTRL,
IWM_CSR_GP_CNTRL_REG_FLAG_MAC_ACCESS_REQ);
/*
* Tell the firmware what we have processed.
* Seems like the hardware gets upset unless we align
* the write by 8??
*/
hw = (hw == 0) ? IWM_RX_RING_COUNT - 1 : hw - 1;
IWM_WRITE(sc, IWM_FH_RSCSR_CHNL0_WPTR, hw & ~7);
}
static void
iwm_intr(void *arg)
{
struct iwm_softc *sc = arg;
int handled = 0;
int r1, r2, rv = 0;
int isperiodic = 0;
IWM_LOCK(sc);
IWM_WRITE(sc, IWM_CSR_INT_MASK, 0);
if (sc->sc_flags & IWM_FLAG_USE_ICT) {
uint32_t *ict = sc->ict_dma.vaddr;
int tmp;
tmp = htole32(ict[sc->ict_cur]);
if (!tmp)
goto out_ena;
/*
* ok, there was something. keep plowing until we have all.
*/
r1 = r2 = 0;
while (tmp) {
r1 |= tmp;
ict[sc->ict_cur] = 0;
sc->ict_cur = (sc->ict_cur+1) % IWM_ICT_COUNT;
tmp = htole32(ict[sc->ict_cur]);
}
/* this is where the fun begins. don't ask */
if (r1 == 0xffffffff)
r1 = 0;
/* i am not expected to understand this */
if (r1 & 0xc0000)
r1 |= 0x8000;
r1 = (0xff & r1) | ((0xff00 & r1) << 16);
} else {
r1 = IWM_READ(sc, IWM_CSR_INT);
/* "hardware gone" (where, fishing?) */
if (r1 == 0xffffffff || (r1 & 0xfffffff0) == 0xa5a5a5a0)
goto out;
r2 = IWM_READ(sc, IWM_CSR_FH_INT_STATUS);
}
if (r1 == 0 && r2 == 0) {
goto out_ena;
}
IWM_WRITE(sc, IWM_CSR_INT, r1 | ~sc->sc_intmask);
/* ignored */
handled |= (r1 & (IWM_CSR_INT_BIT_ALIVE /*| IWM_CSR_INT_BIT_SCD*/));
if (r1 & IWM_CSR_INT_BIT_SW_ERR) {
int i;
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct ieee80211com *ic = &sc->sc_ic;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
struct ieee80211vap *vap = TAILQ_FIRST(&ic->ic_vaps);
The iwm_nic_error() defined if IWM_DEBUG is defined.
2016-04-29 19:10:22 +00:00
#ifdef IWM_DEBUG
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_nic_error(sc);
The iwm_nic_error() defined if IWM_DEBUG is defined.
2016-04-29 19:10:22 +00:00
#endif
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Dump driver status (TX and RX rings) while we're here. */
device_printf(sc->sc_dev, "driver status:\n");
for (i = 0; i < IWM_MVM_MAX_QUEUES; i++) {
struct iwm_tx_ring *ring = &sc->txq[i];
device_printf(sc->sc_dev,
" tx ring %2d: qid=%-2d cur=%-3d "
"queued=%-3d\n",
i, ring->qid, ring->cur, ring->queued);
}
device_printf(sc->sc_dev,
" rx ring: cur=%d\n", sc->rxq.cur);
device_printf(sc->sc_dev,
iwm(4): Don't dereference potentially NULL pointer before NULL check Introduced in r298594. There is no path before the 'vap == NULL' check where vap is not already dereferenced. Reported by: Coverity CID: 1354979 Sponsored by: EMC / Isilon Storage Division
2016-04-26 19:06:28 +00:00
" 802.11 state %d\n", (vap == NULL) ? -1 : vap->iv_state);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
[iwm] restart the VAP if there's a firmware panic. Always print out the firmware panic info before restarting; don't put it behind IWM_DEBUG. Submitted by: bapt Differential Revision: https://reviews.freebsd.org/D6081
2016-04-25 22:00:05 +00:00
/* Don't stop the device; just do a VAP restart */
IWM_UNLOCK(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
[iwm] restart the VAP if there's a firmware panic. Always print out the firmware panic info before restarting; don't put it behind IWM_DEBUG. Submitted by: bapt Differential Revision: https://reviews.freebsd.org/D6081
2016-04-25 22:00:05 +00:00
if (vap == NULL) {
printf("%s: null vap\n", __func__);
return;
}
device_printf(sc->sc_dev, "%s: controller panicked, iv_state = %d; "
"restarting\n", __func__, vap->iv_state);
/* XXX TODO: turn this into a callout/taskqueue */
ieee80211_restart_all(ic);
return;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
if (r1 & IWM_CSR_INT_BIT_HW_ERR) {
handled |= IWM_CSR_INT_BIT_HW_ERR;
device_printf(sc->sc_dev, "hardware error, stopping device\n");
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
iwm_stop(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
rv = 1;
goto out;
}
/* firmware chunk loaded */
if (r1 & IWM_CSR_INT_BIT_FH_TX) {
IWM_WRITE(sc, IWM_CSR_FH_INT_STATUS, IWM_CSR_FH_INT_TX_MASK);
handled |= IWM_CSR_INT_BIT_FH_TX;
sc->sc_fw_chunk_done = 1;
wakeup(&sc->sc_fw);
}
if (r1 & IWM_CSR_INT_BIT_RF_KILL) {
handled |= IWM_CSR_INT_BIT_RF_KILL;
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
if (iwm_check_rfkill(sc)) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev,
"%s: rfkill switch, disabling interface\n",
__func__);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
iwm_stop(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
}
/*
* The Linux driver uses periodic interrupts to avoid races.
* We cargo-cult like it's going out of fashion.
*/
if (r1 & IWM_CSR_INT_BIT_RX_PERIODIC) {
handled |= IWM_CSR_INT_BIT_RX_PERIODIC;
IWM_WRITE(sc, IWM_CSR_INT, IWM_CSR_INT_BIT_RX_PERIODIC);
if ((r1 & (IWM_CSR_INT_BIT_FH_RX | IWM_CSR_INT_BIT_SW_RX)) == 0)
IWM_WRITE_1(sc,
IWM_CSR_INT_PERIODIC_REG, IWM_CSR_INT_PERIODIC_DIS);
isperiodic = 1;
}
if ((r1 & (IWM_CSR_INT_BIT_FH_RX | IWM_CSR_INT_BIT_SW_RX)) || isperiodic) {
handled |= (IWM_CSR_INT_BIT_FH_RX | IWM_CSR_INT_BIT_SW_RX);
IWM_WRITE(sc, IWM_CSR_FH_INT_STATUS, IWM_CSR_FH_INT_RX_MASK);
iwm_notif_intr(sc);
/* enable periodic interrupt, see above */
if (r1 & (IWM_CSR_INT_BIT_FH_RX | IWM_CSR_INT_BIT_SW_RX) && !isperiodic)
IWM_WRITE_1(sc, IWM_CSR_INT_PERIODIC_REG,
IWM_CSR_INT_PERIODIC_ENA);
}
if (__predict_false(r1 & ~handled))
IWM_DPRINTF(sc, IWM_DEBUG_INTR,
"%s: unhandled interrupts: %x\n", __func__, r1);
rv = 1;
out_ena:
iwm_restore_interrupts(sc);
out:
IWM_UNLOCK(sc);
return;
}
/*
* Autoconf glue-sniffing
*/
#define PCI_VENDOR_INTEL 0x8086
#define PCI_PRODUCT_INTEL_WL_3160_1 0x08b3
#define PCI_PRODUCT_INTEL_WL_3160_2 0x08b4
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
#define PCI_PRODUCT_INTEL_WL_3165_1 0x3165
#define PCI_PRODUCT_INTEL_WL_3165_2 0x3166
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
#define PCI_PRODUCT_INTEL_WL_7260_1 0x08b1
#define PCI_PRODUCT_INTEL_WL_7260_2 0x08b2
#define PCI_PRODUCT_INTEL_WL_7265_1 0x095a
#define PCI_PRODUCT_INTEL_WL_7265_2 0x095b
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
#define PCI_PRODUCT_INTEL_WL_8260_1 0x24f3
#define PCI_PRODUCT_INTEL_WL_8260_2 0x24f4
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static const struct iwm_devices {
uint16_t device;
const char *name;
} iwm_devices[] = {
{ PCI_PRODUCT_INTEL_WL_3160_1, "Intel Dual Band Wireless AC 3160" },
{ PCI_PRODUCT_INTEL_WL_3160_2, "Intel Dual Band Wireless AC 3160" },
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
{ PCI_PRODUCT_INTEL_WL_3165_1, "Intel Dual Band Wireless AC 3165" },
{ PCI_PRODUCT_INTEL_WL_3165_2, "Intel Dual Band Wireless AC 3165" },
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
{ PCI_PRODUCT_INTEL_WL_7260_1, "Intel Dual Band Wireless AC 7260" },
{ PCI_PRODUCT_INTEL_WL_7260_2, "Intel Dual Band Wireless AC 7260" },
{ PCI_PRODUCT_INTEL_WL_7265_1, "Intel Dual Band Wireless AC 7265" },
{ PCI_PRODUCT_INTEL_WL_7265_2, "Intel Dual Band Wireless AC 7265" },
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
{ PCI_PRODUCT_INTEL_WL_8260_1, "Intel Dual Band Wireless AC 8260" },
{ PCI_PRODUCT_INTEL_WL_8260_2, "Intel Dual Band Wireless AC 8260" },
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
};
static int
iwm_probe(device_t dev)
{
int i;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
for (i = 0; i < nitems(iwm_devices); i++) {
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (pci_get_vendor(dev) == PCI_VENDOR_INTEL &&
pci_get_device(dev) == iwm_devices[i].device) {
device_set_desc(dev, iwm_devices[i].name);
return (BUS_PROBE_DEFAULT);
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return (ENXIO);
}
static int
iwm_dev_check(device_t dev)
{
struct iwm_softc *sc;
sc = device_get_softc(dev);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sc->sc_hw_rev = IWM_READ(sc, IWM_CSR_HW_REV);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
switch (pci_get_device(dev)) {
case PCI_PRODUCT_INTEL_WL_3160_1:
case PCI_PRODUCT_INTEL_WL_3160_2:
sc->sc_fwname = "iwm3160fw";
sc->host_interrupt_operation_mode = 1;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sc->sc_device_family = IWM_DEVICE_FAMILY_7000;
sc->sc_fwdmasegsz = IWM_FWDMASEGSZ;
return (0);
case PCI_PRODUCT_INTEL_WL_3165_1:
case PCI_PRODUCT_INTEL_WL_3165_2:
sc->sc_fwname = "iwm7265fw";
sc->host_interrupt_operation_mode = 0;
sc->sc_device_family = IWM_DEVICE_FAMILY_7000;
sc->sc_fwdmasegsz = IWM_FWDMASEGSZ;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return (0);
case PCI_PRODUCT_INTEL_WL_7260_1:
case PCI_PRODUCT_INTEL_WL_7260_2:
sc->sc_fwname = "iwm7260fw";
sc->host_interrupt_operation_mode = 1;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sc->sc_device_family = IWM_DEVICE_FAMILY_7000;
sc->sc_fwdmasegsz = IWM_FWDMASEGSZ;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return (0);
case PCI_PRODUCT_INTEL_WL_7265_1:
case PCI_PRODUCT_INTEL_WL_7265_2:
sc->sc_fwname = "iwm7265fw";
sc->host_interrupt_operation_mode = 0;
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sc->sc_device_family = IWM_DEVICE_FAMILY_7000;
sc->sc_fwdmasegsz = IWM_FWDMASEGSZ;
return (0);
case PCI_PRODUCT_INTEL_WL_8260_1:
case PCI_PRODUCT_INTEL_WL_8260_2:
sc->sc_fwname = "iwm8000Cfw";
sc->host_interrupt_operation_mode = 0;
sc->sc_device_family = IWM_DEVICE_FAMILY_8000;
sc->sc_fwdmasegsz = IWM_FWDMASEGSZ_8000;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return (0);
default:
device_printf(dev, "unknown adapter type\n");
return ENXIO;
}
}
static int
iwm_pci_attach(device_t dev)
{
struct iwm_softc *sc;
int count, error, rid;
uint16_t reg;
sc = device_get_softc(dev);
/* Clear device-specific "PCI retry timeout" register (41h). */
reg = pci_read_config(dev, 0x40, sizeof(reg));
pci_write_config(dev, 0x40, reg & ~0xff00, sizeof(reg));
/* Enable bus-mastering and hardware bug workaround. */
pci_enable_busmaster(dev);
reg = pci_read_config(dev, PCIR_STATUS, sizeof(reg));
/* if !MSI */
if (reg & PCIM_STATUS_INTxSTATE) {
reg &= ~PCIM_STATUS_INTxSTATE;
}
pci_write_config(dev, PCIR_STATUS, reg, sizeof(reg));
rid = PCIR_BAR(0);
sc->sc_mem = bus_alloc_resource_any(dev, SYS_RES_MEMORY, &rid,
RF_ACTIVE);
if (sc->sc_mem == NULL) {
device_printf(sc->sc_dev, "can't map mem space\n");
return (ENXIO);
}
sc->sc_st = rman_get_bustag(sc->sc_mem);
sc->sc_sh = rman_get_bushandle(sc->sc_mem);
/* Install interrupt handler. */
count = 1;
rid = 0;
if (pci_alloc_msi(dev, &count) == 0)
rid = 1;
sc->sc_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid, RF_ACTIVE |
(rid != 0 ? 0 : RF_SHAREABLE));
if (sc->sc_irq == NULL) {
device_printf(dev, "can't map interrupt\n");
return (ENXIO);
}
error = bus_setup_intr(dev, sc->sc_irq, INTR_TYPE_NET | INTR_MPSAFE,
NULL, iwm_intr, sc, &sc->sc_ih);
if (sc->sc_ih == NULL) {
device_printf(dev, "can't establish interrupt");
return (ENXIO);
}
sc->sc_dmat = bus_get_dma_tag(sc->sc_dev);
return (0);
}
static void
iwm_pci_detach(device_t dev)
{
struct iwm_softc *sc = device_get_softc(dev);
if (sc->sc_irq != NULL) {
bus_teardown_intr(dev, sc->sc_irq, sc->sc_ih);
bus_release_resource(dev, SYS_RES_IRQ,
rman_get_rid(sc->sc_irq), sc->sc_irq);
pci_release_msi(dev);
}
if (sc->sc_mem != NULL)
bus_release_resource(dev, SYS_RES_MEMORY,
rman_get_rid(sc->sc_mem), sc->sc_mem);
}
static int
iwm_attach(device_t dev)
{
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct iwm_softc *sc = device_get_softc(dev);
struct ieee80211com *ic = &sc->sc_ic;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
int error;
int txq_i, i;
sc->sc_dev = dev;
iwm: reorganize if_iwmvar.h - Change order of data in if_iwmvar.h (like it is in other drivers: defines, data structures, vap/node structures, softc struct and locks); use indentation. - Fix IWM_LOCK(_sc) / IWM_UNLOCK(_sc) macro. - Add IWM_LOCK_INIT / DESTROY(sc) + fix mtx_init() usage. - Wrap iwm_node casts into IWM_NODE() macro. - Drop some fields: * wt_hwqueue from Tx radiotap header; * macaddr[6] from iwm_vap; Approved by: adrian Differential Revision: https://reviews.freebsd.org/D4753
2016-01-03 10:06:10 +00:00
IWM_LOCK_INIT(sc);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
mbufq_init(&sc->sc_snd, ifqmaxlen);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
callout_init_mtx(&sc->sc_watchdog_to, &sc->sc_mtx, 0);
[iwm] add LED blinking for iwm hardware that supports it. Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD, Linux iwlwifi/mvm
2016-06-02 04:42:28 +00:00
callout_init_mtx(&sc->sc_led_blink_to, &sc->sc_mtx, 0);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
TASK_INIT(&sc->sc_es_task, 0, iwm_endscan_cb, sc);
/* PCI attach */
error = iwm_pci_attach(dev);
if (error != 0)
goto fail;
sc->sc_wantresp = -1;
/* Check device type */
error = iwm_dev_check(dev);
if (error != 0)
goto fail;
/*
* We now start fiddling with the hardware
*/
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
/*
* In the 8000 HW family the format of the 4 bytes of CSR_HW_REV have
* changed, and now the revision step also includes bit 0-1 (no more
* "dash" value). To keep hw_rev backwards compatible - we'll store it
* in the old format.
*/
if (sc->sc_device_family == IWM_DEVICE_FAMILY_8000)
sc->sc_hw_rev = (sc->sc_hw_rev & 0xfff0) |
(IWM_CSR_HW_REV_STEP(sc->sc_hw_rev << 2) << 2);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
if (iwm_prepare_card_hw(sc) != 0) {
device_printf(dev, "could not initialize hardware\n");
goto fail;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (sc->sc_device_family == IWM_DEVICE_FAMILY_8000) {
int ret;
uint32_t hw_step;
/*
* In order to recognize C step the driver should read the
* chip version id located at the AUX bus MISC address.
*/
IWM_SETBITS(sc, IWM_CSR_GP_CNTRL,
IWM_CSR_GP_CNTRL_REG_FLAG_INIT_DONE);
DELAY(2);
ret = iwm_poll_bit(sc, IWM_CSR_GP_CNTRL,
IWM_CSR_GP_CNTRL_REG_FLAG_MAC_CLOCK_READY,
IWM_CSR_GP_CNTRL_REG_FLAG_MAC_CLOCK_READY,
25000);
[iwm] Fix iwm_poll_bit() error value check in iwm_attach(). The iwm(4) iwm_poll_bit() function returns 1 on success, and 0 on failure, whereas the iwl_poll_bit() in Linux iwlwifi returns < 0 on failure. So the (ret < 0) check ended up ignoring any error returned by iwm_poll_bit(). Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7932
2016-09-19 22:13:08 +00:00
if (!ret) {
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
device_printf(sc->sc_dev,
"Failed to wake up the nic\n");
goto fail;
}
if (iwm_nic_lock(sc)) {
hw_step = iwm_read_prph(sc, IWM_WFPM_CTRL_REG);
hw_step |= IWM_ENABLE_WFPM;
iwm_write_prph(sc, IWM_WFPM_CTRL_REG, hw_step);
hw_step = iwm_read_prph(sc, IWM_AUX_MISC_REG);
hw_step = (hw_step >> IWM_HW_STEP_LOCATION_BITS) & 0xF;
if (hw_step == 0x3)
sc->sc_hw_rev = (sc->sc_hw_rev & 0xFFFFFFF3) |
(IWM_SILICON_C_STEP << 2);
iwm_nic_unlock(sc);
} else {
device_printf(sc->sc_dev, "Failed to lock the nic\n");
goto fail;
}
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Allocate DMA memory for firmware transfers. */
if ((error = iwm_alloc_fwmem(sc)) != 0) {
device_printf(dev, "could not allocate memory for firmware\n");
goto fail;
}
/* Allocate "Keep Warm" page. */
if ((error = iwm_alloc_kw(sc)) != 0) {
device_printf(dev, "could not allocate keep warm page\n");
goto fail;
}
/* We use ICT interrupts */
if ((error = iwm_alloc_ict(sc)) != 0) {
device_printf(dev, "could not allocate ICT table\n");
goto fail;
}
/* Allocate TX scheduler "rings". */
if ((error = iwm_alloc_sched(sc)) != 0) {
device_printf(dev, "could not allocate TX scheduler rings\n");
goto fail;
}
/* Allocate TX rings */
for (txq_i = 0; txq_i < nitems(sc->txq); txq_i++) {
if ((error = iwm_alloc_tx_ring(sc,
&sc->txq[txq_i], txq_i)) != 0) {
device_printf(dev,
"could not allocate TX ring %d\n",
txq_i);
goto fail;
}
}
/* Allocate RX ring. */
if ((error = iwm_alloc_rx_ring(sc, &sc->rxq)) != 0) {
device_printf(dev, "could not allocate RX ring\n");
goto fail;
}
/* Clear pending interrupts. */
IWM_WRITE(sc, IWM_CSR_INT, 0xffffffff);
ic->ic_softc = sc;
ic->ic_name = device_get_nameunit(sc->sc_dev);
ic->ic_phytype = IEEE80211_T_OFDM; /* not only, but not used */
ic->ic_opmode = IEEE80211_M_STA; /* default to BSS mode */
/* Set device capabilities. */
ic->ic_caps =
IEEE80211_C_STA |
IEEE80211_C_WPA | /* WPA/RSN */
IEEE80211_C_WME |
IEEE80211_C_SHSLOT | /* short slot time supported */
IEEE80211_C_SHPREAMBLE /* short preamble supported */
// IEEE80211_C_BGSCAN /* capable of bg scanning */
;
[iwm] Set full-offload scan flag. Fixes fw panic when already associated. * Starting a scan from wpa_supplicant or via ifconfig while associated, should no longer cause firmware panics or abort early. Tested: * AC7260, STA mode Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D8412
2016-11-01 22:47:58 +00:00
/* Advertise full-offload scanning */
ic->ic_flags_ext = IEEE80211_FEXT_SCAN_OFFLOAD;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
for (i = 0; i < nitems(sc->sc_phyctxt); i++) {
sc->sc_phyctxt[i].id = i;
sc->sc_phyctxt[i].color = 0;
sc->sc_phyctxt[i].ref = 0;
sc->sc_phyctxt[i].channel = NULL;
}
[iwm] fix up RSSI calculations for both scan results and normal RX operations. * hard code a noise floor of -96 for now. The noise floor update code returns some "interesting" values that I can't map to anything useful right now. * Ensure a default noise floor is set - otherwise the initial scan results have a noise floor of '0'. * Fix up the RSSI calculation to be correctly relative to the noise floor. The RSSI routines return an absolute value in dBm - so fix this up. * Cap RSSI values appropriately. * Ensure we pass in a 1/2 dB unit value in to net80211. Tested: * Intel 7260, STA mode iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: hw rev 0x140, fw ver 16.242414.0, address xx:xx:xx:xx:xx:xx
2016-09-18 05:07:18 +00:00
/* Default noise floor */
sc->sc_noise = -96;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Max RSSI */
sc->sc_max_rssi = IWM_MAX_DBM - IWM_MIN_DBM;
[iwm] fix up RSSI calculations for both scan results and normal RX operations. * hard code a noise floor of -96 for now. The noise floor update code returns some "interesting" values that I can't map to anything useful right now. * Ensure a default noise floor is set - otherwise the initial scan results have a noise floor of '0'. * Fix up the RSSI calculation to be correctly relative to the noise floor. The RSSI routines return an absolute value in dBm - so fix this up. * Cap RSSI values appropriately. * Ensure we pass in a 1/2 dB unit value in to net80211. Tested: * Intel 7260, STA mode iwm0: <Intel Dual Band Wireless AC 7260> mem 0xf1400000-0xf1401fff irq 17 at device 0.0 on pci2 iwm0: hw rev 0x140, fw ver 16.242414.0, address xx:xx:xx:xx:xx:xx
2016-09-18 05:07:18 +00:00
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
sc->sc_preinit_hook.ich_func = iwm_preinit;
sc->sc_preinit_hook.ich_arg = sc;
if (config_intrhook_establish(&sc->sc_preinit_hook) != 0) {
device_printf(dev, "config_intrhook_establish failed\n");
goto fail;
}
#ifdef IWM_DEBUG
SYSCTL_ADD_INT(device_get_sysctl_ctx(dev),
SYSCTL_CHILDREN(device_get_sysctl_tree(dev)), OID_AUTO, "debug",
CTLFLAG_RW, &sc->sc_debug, 0, "control debugging");
#endif
IWM_DPRINTF(sc, IWM_DEBUG_RESET | IWM_DEBUG_TRACE,
"<-%s\n", __func__);
return 0;
/* Free allocated memory if something failed during attachment. */
fail:
iwm_detach_local(sc, 0);
return ENXIO;
}
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
static int
iwm_is_valid_ether_addr(uint8_t *addr)
{
char zero_addr[IEEE80211_ADDR_LEN] = { 0, 0, 0, 0, 0, 0 };
if ((addr[0] & 1) || IEEE80211_ADDR_EQ(zero_addr, addr))
return (FALSE);
return (TRUE);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
static int
iwm_update_edca(struct ieee80211com *ic)
{
Convert more ifp->if_softc -> ic->ic_softc. These should be a big no-op. Tested: * make universe
2015-08-17 23:44:38 +00:00
struct iwm_softc *sc = ic->ic_softc;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_printf(sc->sc_dev, "%s: called\n", __func__);
return (0);
}
static void
iwm_preinit(void *arg)
{
struct iwm_softc *sc = arg;
device_t dev = sc->sc_dev;
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct ieee80211com *ic = &sc->sc_ic;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
int error;
IWM_DPRINTF(sc, IWM_DEBUG_RESET | IWM_DEBUG_TRACE,
"->%s\n", __func__);
IWM_LOCK(sc);
if ((error = iwm_start_hw(sc)) != 0) {
device_printf(dev, "could not initialize hardware\n");
IWM_UNLOCK(sc);
goto fail;
}
error = iwm_run_init_mvm_ucode(sc, 1);
iwm_stop_device(sc);
if (error) {
IWM_UNLOCK(sc);
goto fail;
}
device_printf(dev,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
"hw rev 0x%x, fw ver %s, address %s\n",
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
sc->sc_hw_rev & IWM_CSR_HW_REV_TYPE_MSK,
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sc->sc_fwver, ether_sprintf(sc->sc_nvm.hw_addr));
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* not all hardware can do 5GHz band */
if (!sc->sc_nvm.sku_cap_band_52GHz_enable)
memset(&ic->ic_sup_rates[IEEE80211_MODE_11A], 0,
sizeof(ic->ic_sup_rates[IEEE80211_MODE_11A]));
IWM_UNLOCK(sc);
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
iwm_init_channel_map(ic, IEEE80211_CHAN_MAX, &ic->ic_nchans,
ic->ic_channels);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/*
* At this point we've committed - if we fail to do setup,
* we now also have to tear down the net80211 state.
*/
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
ieee80211_ifattach(ic);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ic->ic_vap_create = iwm_vap_create;
ic->ic_vap_delete = iwm_vap_delete;
ic->ic_raw_xmit = iwm_raw_xmit;
ic->ic_node_alloc = iwm_node_alloc;
ic->ic_scan_start = iwm_scan_start;
ic->ic_scan_end = iwm_scan_end;
ic->ic_update_mcast = iwm_update_mcast;
iwm: switch to ieee80211_add_channel() - Switch to ieee80211_add_channel(). - Parse/set more channel flags (DFS, NOADHOC). - Add ic_getradiocaps() method. Reviewed by: adrian Differential Revision: https://reviews.freebsd.org/D6150
2016-05-01 14:27:27 +00:00
ic->ic_getradiocaps = iwm_init_channel_map;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ic->ic_set_channel = iwm_set_channel;
ic->ic_scan_curchan = iwm_scan_curchan;
ic->ic_scan_mindwell = iwm_scan_mindwell;
ic->ic_wme.wme_update = iwm_update_edca;
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
ic->ic_parent = iwm_parent;
ic->ic_transmit = iwm_transmit;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_radiotap_attach(sc);
if (bootverbose)
ieee80211_announce(ic);
IWM_DPRINTF(sc, IWM_DEBUG_RESET | IWM_DEBUG_TRACE,
"<-%s\n", __func__);
config_intrhook_disestablish(&sc->sc_preinit_hook);
return;
fail:
config_intrhook_disestablish(&sc->sc_preinit_hook);
iwm_detach_local(sc, 0);
}
/*
* Attach the interface to 802.11 radiotap.
*/
static void
iwm_radiotap_attach(struct iwm_softc *sc)
{
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
struct ieee80211com *ic = &sc->sc_ic;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_DPRINTF(sc, IWM_DEBUG_RESET | IWM_DEBUG_TRACE,
"->%s begin\n", __func__);
ieee80211_radiotap_attach(ic,
&sc->sc_txtap.wt_ihdr, sizeof(sc->sc_txtap),
IWM_TX_RADIOTAP_PRESENT,
&sc->sc_rxtap.wr_ihdr, sizeof(sc->sc_rxtap),
IWM_RX_RADIOTAP_PRESENT);
IWM_DPRINTF(sc, IWM_DEBUG_RESET | IWM_DEBUG_TRACE,
"->%s end\n", __func__);
}
static struct ieee80211vap *
iwm_vap_create(struct ieee80211com *ic, const char name[IFNAMSIZ], int unit,
enum ieee80211_opmode opmode, int flags,
const uint8_t bssid[IEEE80211_ADDR_LEN],
const uint8_t mac[IEEE80211_ADDR_LEN])
{
struct iwm_vap *ivp;
struct ieee80211vap *vap;
if (!TAILQ_EMPTY(&ic->ic_vaps)) /* only one at a time */
return NULL;
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
ivp = malloc(sizeof(struct iwm_vap), M_80211_VAP, M_WAITOK | M_ZERO);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
vap = &ivp->iv_vap;
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
ieee80211_vap_setup(ic, vap, name, unit, opmode, flags, bssid);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
vap->iv_bmissthreshold = 10; /* override default */
/* Override with driver methods. */
ivp->iv_newstate = vap->iv_newstate;
vap->iv_newstate = iwm_newstate;
ieee80211_ratectl_init(vap);
/* Complete setup. */
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
ieee80211_vap_attach(vap, iwm_media_change, ieee80211_media_status,
mac);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
ic->ic_opmode = opmode;
return vap;
}
static void
iwm_vap_delete(struct ieee80211vap *vap)
{
struct iwm_vap *ivp = IWM_VAP(vap);
ieee80211_ratectl_deinit(vap);
ieee80211_vap_detach(vap);
free(ivp, M_80211_VAP);
}
static void
iwm_scan_start(struct ieee80211com *ic)
{
struct ieee80211vap *vap = TAILQ_FIRST(&ic->ic_vaps);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
struct iwm_softc *sc = ic->ic_softc;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
int error;
IWM_LOCK(sc);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (isset(sc->sc_enabled_capa, IWM_UCODE_TLV_CAPA_UMAC_SCAN))
error = iwm_mvm_umac_scan(sc);
else
error = iwm_mvm_lmac_scan(sc);
if (error != 0) {
[iwm] Clean up iwm(4) scanning logic a bit. Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD 8f3ffab9136e33263d424275ec28f57ad2096437
2016-06-02 04:54:56 +00:00
device_printf(sc->sc_dev, "could not initiate 2 GHz scan\n");
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_UNLOCK(sc);
ieee80211_cancel_scan(vap);
[iwm] add LED blinking for iwm hardware that supports it. Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD, Linux iwlwifi/mvm
2016-06-02 04:42:28 +00:00
} else {
iwm_led_blink_start(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
IWM_UNLOCK(sc);
[iwm] add LED blinking for iwm hardware that supports it. Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD, Linux iwlwifi/mvm
2016-06-02 04:42:28 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
static void
iwm_scan_end(struct ieee80211com *ic)
{
[iwm] add LED blinking for iwm hardware that supports it. Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD, Linux iwlwifi/mvm
2016-06-02 04:42:28 +00:00
struct ieee80211vap *vap = TAILQ_FIRST(&ic->ic_vaps);
struct iwm_softc *sc = ic->ic_softc;
IWM_LOCK(sc);
iwm_led_blink_stop(sc);
if (vap->iv_state == IEEE80211_S_RUN)
iwm_mvm_led_enable(sc);
IWM_UNLOCK(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
}
static void
iwm_update_mcast(struct ieee80211com *ic)
{
}
static void
iwm_set_channel(struct ieee80211com *ic)
{
}
static void
iwm_scan_curchan(struct ieee80211_scan_state *ss, unsigned long maxdwell)
{
}
static void
iwm_scan_mindwell(struct ieee80211_scan_state *ss)
{
return;
}
void
iwm_init_task(void *arg1)
{
struct iwm_softc *sc = arg1;
IWM_LOCK(sc);
while (sc->sc_flags & IWM_FLAG_BUSY)
msleep(&sc->sc_flags, &sc->sc_mtx, 0, "iwmpwr", 0);
sc->sc_flags |= IWM_FLAG_BUSY;
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
iwm_stop(sc);
if (sc->sc_ic.ic_nrunning > 0)
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
iwm_init(sc);
sc->sc_flags &= ~IWM_FLAG_BUSY;
wakeup(&sc->sc_flags);
IWM_UNLOCK(sc);
}
static int
iwm_resume(device_t dev)
{
[iwm] implement suspend/resume through ieee80211_{suspend,resume}_all This allows wifi to associate correctly after a suspend/resume cycle. Yes, I'm using this now day to day. Tested: * Intel 7260AC, STA mode
2016-04-26 04:40:59 +00:00
struct iwm_softc *sc = device_get_softc(dev);
int do_reinit = 0;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
uint16_t reg;
/* Clear device-specific "PCI retry timeout" register (41h). */
reg = pci_read_config(dev, 0x40, sizeof(reg));
pci_write_config(dev, 0x40, reg & ~0xff00, sizeof(reg));
iwm_init_task(device_get_softc(dev));
[iwm] implement suspend/resume through ieee80211_{suspend,resume}_all This allows wifi to associate correctly after a suspend/resume cycle. Yes, I'm using this now day to day. Tested: * Intel 7260AC, STA mode
2016-04-26 04:40:59 +00:00
IWM_LOCK(sc);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (sc->sc_flags & IWM_FLAG_SCANNING) {
sc->sc_flags &= ~IWM_FLAG_SCANNING;
[iwm] implement suspend/resume through ieee80211_{suspend,resume}_all This allows wifi to associate correctly after a suspend/resume cycle. Yes, I'm using this now day to day. Tested: * Intel 7260AC, STA mode
2016-04-26 04:40:59 +00:00
do_reinit = 1;
}
IWM_UNLOCK(sc);
if (do_reinit)
ieee80211_resume_all(&sc->sc_ic);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return 0;
}
static int
iwm_suspend(device_t dev)
{
[iwm] implement suspend/resume through ieee80211_{suspend,resume}_all This allows wifi to associate correctly after a suspend/resume cycle. Yes, I'm using this now day to day. Tested: * Intel 7260AC, STA mode
2016-04-26 04:40:59 +00:00
int do_stop = 0;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
struct iwm_softc *sc = device_get_softc(dev);
[iwm] implement suspend/resume through ieee80211_{suspend,resume}_all This allows wifi to associate correctly after a suspend/resume cycle. Yes, I'm using this now day to day. Tested: * Intel 7260AC, STA mode
2016-04-26 04:40:59 +00:00
do_stop = !! (sc->sc_ic.ic_nrunning > 0);
ieee80211_suspend_all(&sc->sc_ic);
if (do_stop) {
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
IWM_LOCK(sc);
iwm_stop(sc);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
sc->sc_flags |= IWM_FLAG_SCANNING;
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
IWM_UNLOCK(sc);
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return (0);
}
static int
iwm_detach_local(struct iwm_softc *sc, int do_net80211)
{
iwm: revert r293178 This optimization is not proper (and causes kernel panic), since driver checks fw_status to optimize away parsing stage if it was already done. Reported by: dchagin
2016-01-05 20:09:26 +00:00
struct iwm_fw_info *fw = &sc->sc_fw;
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
device_t dev = sc->sc_dev;
int i;
iwm: If firmware load fails during init via iwm_preinit() we bail out before calling ieee80211_ifattach() so the taskqueue hasn't been initialized. Don't try to drain it, we'll panic. Looks like this issue was introduced in r303326. Reviewed by: avos, sbruno, adrian Differential Revision: https://reviews.freebsd.org/D8499
2016-11-12 18:33:56 +00:00
if (do_net80211)
ieee80211_draintask(&sc->sc_ic, &sc->sc_es_task);
iwm(4): switch to ieee80211_runtask() Submitted by: Andiry Voskoboinyk (s3erios@gmail.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D5054
2016-07-25 23:44:44 +00:00
[iwm] add LED blinking for iwm hardware that supports it. Submitted by: Imre Vadasz <imre@vdsz.com> Obtained from: DragonflyBSD, Linux iwlwifi/mvm
2016-06-02 04:42:28 +00:00
callout_drain(&sc->sc_led_blink_to);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
callout_drain(&sc->sc_watchdog_to);
iwm_stop_device(sc);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
if (do_net80211) {
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
ieee80211_ifdetach(&sc->sc_ic);
iwm(4) synchronize driver to DragonFlyBSD version and recent f/w update. Submitted by: Kevin Bowling (kevin.bowling@kev009.com) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D6967
2016-07-26 00:02:17 +00:00
}
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
[iwm] Add and use iwm_phy_db_free(), to plug phy_db memory leak. (Together with other iwm(4) memory leak fixes) Memory leakage in M_DEVBUF is now at ca. 2KB for each iwm(4) module load/unload cycle. Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Obtained from: DragonflyBSD git eaf551a1d464c643e98ce5781971dd32124e9af1 Differential Revision: https://reviews.freebsd.org/D6819
2016-06-23 01:11:52 +00:00
iwm_phy_db_free(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Free descriptor rings */
[iwm] free RX ring / NVM memory after they're used. * Free RX ring during detach * Free NVM memory after parsing Tested: * 7260, STA mode Submitted by: Imre Vadasz <imre@vdsz.com> Approved by: re (gjb) Obtained from: dragonflybsd Differential Revision: https://reviews.freebsd.org/D6817
2016-06-16 17:59:15 +00:00
iwm_free_rx_ring(sc, &sc->rxq);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
for (i = 0; i < nitems(sc->txq); i++)
iwm_free_tx_ring(sc, &sc->txq[i]);
iwm: revert r293178 This optimization is not proper (and causes kernel panic), since driver checks fw_status to optimize away parsing stage if it was already done. Reported by: dchagin
2016-01-05 20:09:26 +00:00
/* Free firmware */
if (fw->fw_fp != NULL)
iwm_fw_info_free(fw);
iwm: free firmware related resources after uploading it to the hardware iwn(4) / wpi(4) works in the same way (read_firmware() -> hw_init() -> firmware_put()) Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D4766
2016-01-04 21:07:08 +00:00
/* Free scheduler */
[iwm] Remove wrappers around iwm_dma_contig_free() calls. No functional changes. Inspired by: OpenBSD Approved by: adrian (mentor) Differential Revision: https://reviews.freebsd.org/D7933
2016-09-19 22:16:03 +00:00
iwm_dma_contig_free(&sc->sched_dma);
iwm_dma_contig_free(&sc->ict_dma);
iwm_dma_contig_free(&sc->kw_dma);
iwm_dma_contig_free(&sc->fw_dma);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
/* Finished with the hardware - detach things */
iwm_pci_detach(dev);
Replay r286410. Change KPI of how device drivers that provide wireless connectivity interact with the net80211 stack. Historical background: originally wireless devices created an interface, just like Ethernet devices do. Name of an interface matched the name of the driver that created. Later, wlan(4) layer was introduced, and the wlanX interfaces become the actual interface, leaving original ones as "a parent interface" of wlanX. Kernelwise, the KPI between net80211 layer and a driver became a mix of methods that pass a pointer to struct ifnet as identifier and methods that pass pointer to struct ieee80211com. From user point of view, the parent interface just hangs on in the ifconfig list, and user can't do anything useful with it. Now, the struct ifnet goes away. The struct ieee80211com is the only KPI between a device driver and net80211. Details: - The struct ieee80211com is embedded into drivers softc. - Packets are sent via new ic_transmit method, which is very much like the previous if_transmit. - Bringing parent up/down is done via new ic_parent method, which notifies driver about any changes: number of wlan(4) interfaces, number of them in promisc or allmulti state. - Device specific ioctls (if any) are received on new ic_ioctl method. - Packets/errors accounting are done by the stack. In certain cases, when driver experiences errors and can not attribute them to any specific interface, driver updates ic_oerrors or ic_ierrors counters. Details on interface configuration with new world order: - A sequence of commands needed to bring up wireless DOESN"T change. - /etc/rc.conf parameters DON'T change. - List of devices that can be used to create wlan(4) interfaces is now provided by net.wlan.devices sysctl. Most drivers in this change were converted by me, except of wpi(4), that was done by Andriy Voskoboinyk. Big thanks to Kevin Lo for testing changes to at least 8 drivers. Thanks to pluknet@, Oliver Hartmann, Olivier Cochard, gjb@, mmoll@, op@ and lev@, who also participated in testing. Reviewed by: adrian Sponsored by: Netflix Sponsored by: Nginx, Inc.
2015-08-27 08:56:39 +00:00
mbufq_drain(&sc->sc_snd);
iwm: reorganize if_iwmvar.h - Change order of data in if_iwmvar.h (like it is in other drivers: defines, data structures, vap/node structures, softc struct and locks); use indentation. - Fix IWM_LOCK(_sc) / IWM_UNLOCK(_sc) macro. - Add IWM_LOCK_INIT / DESTROY(sc) + fix mtx_init() usage. - Wrap iwm_node casts into IWM_NODE() macro. - Drop some fields: * wt_hwqueue from Tx radiotap header; * macaddr[6] from iwm_vap; Approved by: adrian Differential Revision: https://reviews.freebsd.org/D4753
2016-01-03 10:06:10 +00:00
IWM_LOCK_DESTROY(sc);
Import OpenBSD's iwm WiFi driver for Intel 3160/7260/7265. There are still several bugs, but I've been using it for a while now. Thanks to all the testers and to Adrian for his help with this driver. This driver isn't connected to the build yet, but it will be soon. There's no MFC planned because the driver isn't very stable yet. Reviewed by: adrian Obtained from: https://github.com/rpaulo/iwm Tested by: adrian, gjb, dumbbell (others that I forgot). Relnotes: yes
2015-08-08 06:06:48 +00:00
return (0);
}
static int
iwm_detach(device_t dev)
{
struct iwm_softc *sc = device_get_softc(dev);
return (iwm_detach_local(sc, 1));
}
static device_method_t iwm_pci_methods[] = {
/* Device interface */
DEVMETHOD(device_probe, iwm_probe),
DEVMETHOD(device_attach, iwm_attach),
DEVMETHOD(device_detach, iwm_detach),
DEVMETHOD(device_suspend, iwm_suspend),
DEVMETHOD(device_resume, iwm_resume),
DEVMETHOD_END
};
static driver_t iwm_pci_driver = {
"iwm",
iwm_pci_methods,
sizeof (struct iwm_softc)
};
static devclass_t iwm_devclass;
DRIVER_MODULE(iwm, pci, iwm_pci_driver, iwm_devclass, NULL, NULL);
MODULE_DEPEND(iwm, firmware, 1, 1, 1);
MODULE_DEPEND(iwm, pci, 1, 1, 1);
MODULE_DEPEND(iwm, wlan, 1, 1, 1);
Reference in New Issue Copy Permalink