Commit Graph

191 Commits

Author SHA1 Message Date
Coleman Kane
e0d12c4b94 Fix regression in last set of commits. Submitted via e-mail and then
nagged again via PR. Thank Paul for his persistence and contributions.

PR:		136895
Submitted by:	Paul B. Mahol <onemda@gmail.com>
Reviewed by:	sam (timeout, 10 days), weongyo (timeout, 10 days), me
Approved by:	re (Kostik Belousov <kostikbel@gmail.com>)
2009-07-20 23:21:19 +00:00
Robert Watson
eb956cd041 Use if_maddr_rlock()/if_maddr_runlock() rather than IF_ADDR_LOCK()/
IF_ADDR_UNLOCK() across network device drivers when accessing the
per-interface multicast address list, if_multiaddrs.  This will
allow us to change the locking strategy without affecting our driver
programming interface or binary interface.

For two wireless drivers, remove unnecessary locking, since they
don't actually access the multicast address list.

Approved by:	re (kib)
MFC after:	6 weeks
2009-06-26 11:45:06 +00:00
Coleman Kane
7c1a5c7c93 Code cleanup by moving some repetitive code into an ndis_get_bssid_list
helper function. Also, add ieee80211_announce() call for bootverbose case.

Submitted by:	Paul B. Mahol <onemda@gmail.com>
2009-06-23 14:37:07 +00:00
Andrew Thompson
ed6d949afd - Make struct usb_xfer opaque so that drivers can not access the internals
- Reduce the number of headers needed for a usb driver, the common case is just   usb.h and usbdi.h
2009-06-23 02:19:59 +00:00
Coleman Kane
d9cacbcf4f Replace use of ic->ic_flags with vap->iv_flags to operate on per-vap flags
for ndis 802.11 work.

Submitted by:	Paul B. Mahol <onemda@gmail.com>
2009-06-18 11:12:10 +00:00
Andrew Thompson
a593f6b8de s/usb2_/usb_|usbd_/ on all function names for the USB stack. 2009-06-15 01:02:43 +00:00
Coleman Kane
badff0e4d8 Fix an LOR
Submitted by:	Paul B. Mahol <onemda@gmail.com>
MFC after:	1 week
2009-06-10 18:00:10 +00:00
Andrew Thompson
5cdee58296 Commit missed driver_info to driver_ivar change in usb_attach_args.
Pointed out by:	kib
2009-06-08 19:56:23 +00:00
Sam Leffler
e6c3129877 fix setting of ni_txrate
Submitted by:	"Paul B. Mahol" <onemda@gmail.com>
2009-06-02 20:32:13 +00:00
Andrew Thompson
e0a69b51ac s/usb2_/usb_/ on all typedefs for the USB stack. 2009-05-29 18:46:57 +00:00
Andrew Thompson
760bc48e7e s/usb2_/usb_/ on all C structs for the USB stack. 2009-05-28 17:36:36 +00:00
Andrew Thompson
fadc970b77 Fix a few variable renames of usb2_mode outside dev/usb. 2009-05-21 02:09:12 +00:00
Sam Leffler
74904f7d58 update for net80211 rx api change 2009-05-20 23:33:40 +00:00
Andrew Thompson
5efea30f03 Create a taskqueue for each wireless interface which provides a serialised
sleepable context for net80211 driver callbacks. This removes the need for USB
and firmware based drivers to roll their own code to defer the chip programming
for state changes, scan requests, channel changes and mcast/promisc updates.
When a driver callback completes the hardware state is now guaranteed to have
been updated and is in sync with net80211 layer.

This nukes around 1300 lines of code from the wireless device drivers making
them more readable and less race prone.

The net80211 layer has been updated as follows
 - all state/channel changes are serialised on the taskqueue.
 - ieee80211_new_state() always queues and can now be called from any context
 - scanning runs from a single taskq function and executes to completion. driver
   callbacks are synchronous so the channel, phy mode and rx filters are
   guaranteed to be set in hardware before probe request frames are
   transmitted.

Help and contributions from Sam Leffler.

Reviewed by:	sam
2009-05-02 15:14:18 +00:00
Andrew Thompson
a0c4317cc2 Only call ieee80211_start_all() if its a 80211 device.
Submitted by:	Paul B. Mahol
2009-04-16 23:42:00 +00:00
Sam Leffler
9fb0fccb89 o add a capability for drivers that require 802.3 encapsulation of
frames passed down through the transmit path
o mark ndis requiring 802.3 encap'd frames

Reviewed by:	"Paul B. Mahol" <onemda@gmail.com>, thompsa
2009-04-08 16:19:50 +00:00
Sam Leffler
29aca94080 Eliminate ic_myaddr so changing the mac address of a device works correctly:
o remove ic_myaddr from ieee80211com
o change ieee80211_ifattach to take the mac address of the physical device
  and use that to setup the lladdr.
o replace all references to ic_myaddr in drivers by IF_LLADDR
o related cleanups (e.g. kill dead code)

PR:		kern/133178
Reviewed by:	thompsa, rpaulo
2009-03-29 17:59:14 +00:00
Weongyo Jeong
c4f82f895d set NULL after free to avoid duplicate free.
Tested by:	Ganbold <ganbold_at_micom.mng.net>
2009-03-24 04:20:17 +00:00
Weongyo Jeong
577b9fa3f8 Some NDIS USB drivers try to call URB funcs like URB_FUNCTION_VENDOR_xxx
or URB_FUNCTION_CLASS_xxx with HAL preemption lock that means it's
non-sleepable during USB requests though usb2_do_request() requires a
sleep so it needs to send queries to the default pipe without those
interfaces to avoid sleep.
2009-03-18 02:38:35 +00:00
Weongyo Jeong
2c964f43b6 o change a lock model based on HAL preemption lock to a normal mtx.
Based on the HAL preemption lock there is a problem on SMP machines
  and causes a panic.
o When a device detached the current tactic to detach NDIS USB driver is
  to call SURPRISE_REMOVED event.  So it don't need to call
  ndis_halt_nic() again.  This fixes some page faults when some drivers
  work abnormal.
o it assumes now that URB_FUNCTION_BULK_OR_INTERRUPT_TRANSFER is in
  DISPATCH_LEVEL (non-sleepable) and as further work
  URB_FUNCTION_VENDOR_XXX and URB_FUNCTION_CLASS_XXX should be.

Reviewed by:	Hans Petter Selasky <hselasky_at_freebsd.org>
Tested by:	Paul B. Mahol <onemda_at_gmail.com>
2009-03-12 02:51:55 +00:00
Warner Losh
155a83e87a remove now-redunant cardbus attachment. 2009-03-09 13:23:54 +00:00
Sam Leffler
f11fe89814 Fix TXPMGT handling:
o correct dBm<->mW conversion logic
o set net80211 TXPMGT capability only if driver reports it is capable

PR:		kern/132342
Submitted by:	"Paul B. Mahol" <onemda@gmail.com>
2009-03-09 02:37:52 +00:00
Weongyo Jeong
6affafd098 o port NDIS USB support from USB1 to the new usb(USB2).
o implement URB_FUNCTION_ABORT_PIPE handling.
o remove unused code related with canceling the timer list for USB
  drivers.
o whitespace cleanup and style(9)

Obtained from:	hps's original patch
2009-03-07 07:26:22 +00:00
Andrew Thompson
3975e3a1ea Move usb to a graveyard location under sys/legacy/dev, it is intended that the
new USB2 stack will fully replace this for 8.0.

Remove kernel modules, a subsequent commit will update conf/files. Unhook
usbdevs from the build.
2009-02-23 18:16:17 +00:00
Andrew Thompson
1e736167a5 Provide a NDIS tx power level converstion to/from net80211, this uses a loookup
table to approximate mW<->dBM.

Submitted by:	Paul B. Mahol
2009-01-12 17:15:09 +00:00
Weongyo Jeong
c862b05ba7 fix a typo to set the 48Mbps data rate.
PR:		kern/130189
Submitted by:	Paul B. Mahol <onemda_at_gmail.com>
Reviewed by:	sam
MFC after:	3 weeks
2009-01-09 02:31:51 +00:00
Weongyo Jeong
b3974c00b5 Integrate the NDIS USB support code to CURRENT.
Now the NDISulator supports NDIS USB drivers that it've tested with
devices as follows:

  - Anygate XM-142 (Conexant)
  - Netgear WG111v2 (Realtek)
  - U-Khan UW-2054u (Marvell)
  - Shuttle XPC Accessory PN20 (Realtek)
  - ipTIME G054U2 (Ralink)
  - UNiCORN WL-54G (ZyDAS)
  - ZyXEL G-200v2 (ZyDAS)

All of them succeeded to attach and worked though there are still some
problems that it's expected to be solved.

To use NDIS USB support, you should rebuild and install ndiscvt(8) and
if you encounter a problem to attach please set `hw.ndisusb.halt' to
0 then retry.

I expect no changes of the NDIS code for PCI, PCMCIA devices.

Obtained from:  //depot/projects/ndisusb/...
2008-12-27 08:03:32 +00:00
Sam Leffler
e0a11d8bec hookup mcast and promisc callbacks
Submitted by:	"Paul B. Mahol" <onemda@gmail.com>
2008-11-30 20:48:57 +00:00
Andrew Thompson
a1f8daf412 Fake the assoc id so that ndis can work on the latest net80211.
PR:		kern/128750
Submitted by:	Paul B. Mahol
2008-11-11 03:36:15 +00:00
Weongyo Jeong
db102f9a37 sc->ndis_txidx should be cycle between 0 and sc->ndis_maxpkts, not
NDIS_TXPKTS and don't allocate unused extra spaces for sc->ndis_txarray
and sc->ndis_txpool.

PR:		kern/127644
Submitted by:	Antoine Pelisse <apelisse_at_gmail.com>
MFC after:	1 week
2008-10-04 04:15:39 +00:00
Coleman Kane
093d7be50e Fix a mutex LOR introduced by the conversion of if_ndis from spinlocks to
mutexes and replacing the obsolete if_watchdog interface. The ndis_ticktask
function calls into ieee80211_new_state under one condition with NDIS_LOCK
held. The ieee80211_new_state would call into ndis_start in some cases too,
resulting in the occasional case where ndis_start acquires NDIS_LOCK from
inside the NDIS_LOCK held by ndis_ticktask.

Obtained from:	Paul B. Mahol <onemda@gmail.com>
MFC after:	1 week
2008-07-09 00:10:55 +00:00
Coleman Kane
21a6592999 Convert ndis_spinlock to ndis_mtx and start using the sleepable
mtx interface for NDIS_LOCK/UNLOCK. This should result in less
CPU utilization on behalf of the ndis driver. Additionally, this
commit also fixes a potential LOR in the ndis_tick code, by
not locking inside the ndis_tick function, but instead delegating
that work to the helpers called through IoQueueWorkItem. The
way that this is currently set up for NDIS prevents us from
simply implementing a callout_init_mtx mechanism.

However, the helper functions that handle the various timeout
cases implement fine-grained locking using the spinlocks provided
by the NDIS-compat layer, and using the mtx that is added with
this commit. This leaves the following ndis_softc members operated
on in ndis_tick in an unlocked context:

  * ndis_hang_timer - Only modified outside of ndis_tick once, before
                      the first callout_reset to schedule ndis_tick
  * ifp->if_oerrors - Only incremented in two places, which should be
                      an atomic op
  * ndis_tx_timer   - Assigned to 5 (when guaranteed to be 0) or 0
                      (in txeof), to indicate to ndis_tick what to
                      do. This is the only member of which I was
                      suspicious for needing the NDIS_LOCK here. My
                      testing (and another's) have been fine so far.
  * ndis_stat_callout - Only uses a simple set of callout routines,
                        callout_reset only called by ndis_tick after
                        the initial reset, and then callout_drain is
                        used exactly once in shutdown code.

The benefit is that ndis_tick doesn't acquire NDIS_LOCK unless one of
the timeout conditions is flagged, and it still obeys the locking
order semantics that are dictated by the NDIS layer at the moment. I
have been investigating a more thorough s/spinlock/mtx/ of the NDIS
layer, but the simplest naive approach (replace KeAcquireSpinLock
with an mtx implementation) has anti-succeeded for me so far. This
is a good first step though.

Tested by:	onemda@gmail.com
Reviewed by:	current@, jhb, thompsa
Proposed by:	jhb
2008-06-11 13:40:15 +00:00
Coleman Kane
e97f228a10 Update if_ndis to remove the legacy if_watchdog interface and
bring it more up to date. The watchdog timer, and its
associated code, is all collapsed into the ndis_tick function
that was implemented for the NDIS-subsystem watchdog. This
implementation is similar to what numerous other drivers use
to implement the watchdog.

Reviewed by:	thompsa, jhb
MFC after:	2 weeks
2008-06-03 00:55:48 +00:00
Weongyo Jeong
e54b92f37b Fix a panic that it's occurred when NDIS init handler returned a error
by unknown reasons.   In this case, sc->ifp is still NULL so it will
cause a page fault during calling ndis_detach()
2008-05-30 07:17:51 +00:00
Sam Leffler
c43feede8b Minor cleanup of vap create work:
o add IEEE80211_C_STA capability to indicate sta mode is supported
  (was previously assumed) and mark drivers as capable
o add ieee80211_opcap array to map an opmode to the equivalent capability bit
o move IEEE80211_C_OPMODE definition to where capabilities are defined so it's
  clear it should be kept in sync (on future additions)
o check device capabilities in clone create before trying to create a vap;
  this makes driver checks unneeded
o make error codes return on failed clone request unique
o temporarily add console printfs on clone request failures to aid in
  debugging; these will move under DIAGNOSTIC or similar before release
2008-05-12 00:15:30 +00:00
Andrew Thompson
6ba643ce87 - Associate from a taskq as we can deadlock on the ndis hal and the com lock.
- Remove double vap init (ieee80211_start_all)
- Keep ic_curchan in sync with the scan results.
2008-05-10 20:12:43 +00:00
Andrew Thompson
517b468a5d Partially revert the last rev. Do call ndis_setstate_80211() when we up the
interface but break out the associate code into a separate function. This fixes
association with an 11b Apple Airport.

Reported by:	Ted Lindgreen
2008-05-10 20:07:00 +00:00
Andrew Thompson
3fec7550c6 Do not call ndis_setstate_80211() until we are ready to associate, the vap may
not have been created yet and will panic. This requires ndis_scan() to always
set the SSID.

Reported by:	Ben Kaduk
2008-05-01 05:11:33 +00:00
Andrew Thompson
31a8c1edd8 Unify all the wifi *_ioctl routines
- Limit grabbing the lock to SIOCSIFFLAGS.
 - Move ieee80211_start_all() to SIOCSIFFLAGS.
 - Remove SIOCSIFMEDIA as it is not useful.
 - Limit ether_ioctl to only SIOCGIFADDR. SIOCSIFADDR and SIOCSIFMTU have no
   affect as there is no input/output path in the vap parent.  The vap code
   will handle the reinit of the mac address changes.
 - Split off ndis_ioctl_80211 as it was getting too different to wired devices.

This fixes a copyout while locked and a lock recursion.

Reviewed by:		sam
2008-05-01 04:55:00 +00:00
Sam Leffler
b032f27c36 Multi-bss (aka vap) support for 802.11 devices.
Note this includes changes to all drivers and moves some device firmware
loading to use firmware(9) and a separate module (e.g. ral).  Also there
no longer are separate wlan_scan* modules; this functionality is now
bundled into the wlan module.

Supported by:	Hobnob and Marvell
Reviewed by:	many
Obtained from:	Atheros (some bits)
2008-04-20 20:35:46 +00:00
Coleman Kane
12b6a0f87c Change the 1 in callout_init to CALLOUT_MPSAFE. I made the change in my
tree, but never committed it into git before making my patch!

Submitted by:	thompsa
MFC after:	1 month
2008-04-17 22:57:33 +00:00
Coleman Kane
aba7c5b84b Change the timeout(9) usage in if_ndis to a callout(9) implementation,
as the former is becoming deprecated and exhibits some extraneous
Giant-locking. The new callout(9) is declared MPSAFE, so it may
improve concurrency.

Tested by:	matteo
Silence from:	wpaul
MFC after:	1 month
2008-04-17 22:01:38 +00:00
John Baldwin
9a57f6280c Use pause(9) instead of a tsleep() on a stack variable.
MFC after:	1 week
2008-01-17 20:11:58 +00:00
Andrew Thompson
2aec24b515 sc->ndis_tq variable is only initialized when a driver module is for wireless
NICs.

PR:		kern/118439
Submitted by:	Weongyo Jeong
2007-12-09 20:14:01 +00:00
Andrew Thompson
30d9c7698a Preallocate some space to return the scan results, some drivers do not properly
pass back the desired buffer length. This fixes scanning with the Marvell
88W8335 and BCM4328 wireless cards.

PR:		kern/118370
Submitted by:	Weongyo Jeong
Tested by:	Ed Schouten
2007-12-07 10:37:25 +00:00
Wojciech A. Koszek
8b249eccbc Cast address of 'bssid' to uint8_t since printf() expects to get unsigned
argument in %6D case.

Submitted by:	thompsa
Compile-tested:	wkoszek
Reviewed by:	cognet (mentor)
Approved by:	cognet (mentor)
2007-12-04 20:48:32 +00:00
Andrew Thompson
7218dd5f5a Remove compatibilty defines to simplify reading the code, this is around 10% of
the total LOC. These are mostly for the 6.x branch and it will be easier to
cherrypick any changes (if at all).
2007-12-03 21:51:28 +00:00
Andrew Thompson
4d50f65ceb Hide a debug printf, NDIS_DEBUG is always defined and we test the sysctl.
MFC after:	3 days
2007-12-02 09:03:43 +00:00
Julian Elischer
3745c395ec Rename the kthread_xxx (e.g. kthread_create()) calls
to kproc_xxx as they actually make whole processes.
Thos makes way for us to add REAL kthread_create() and friends
that actually make theads. it turns out that most of these
calls actually end up being moved back to the thread version
when it's added. but we need to make this cosmetic change first.

I'd LOVE to do this rename in 7.0  so that we can eventually MFC the
new kthread_xxx() calls.
2007-10-20 23:23:23 +00:00
Andrew Thompson
c4dd9fb67a Fix up ndis interaction with net80211
- make NDIS_DEBUG a sysctl
 - default to IEEE80211_MODE_11B if the card doesnt tell us the channels
 - dont mess with ic_des_chan when we assosciate
 - Allow a directed scan by setting the ESSID before scanning (verified
   with wireshark). Hidden APs probably wouldnt have worked before.
 - Grab the channel type and use it to look up the correct curchan for
   the scan results (mistakenly used 11B before)
 - Fix memory leak in the ndis_scan_results

Tested by:	matteo
Reviewed by:	sam
Approved by:	re (rwatson)
2007-07-26 20:11:16 +00:00