freebsd-nq/sys/dev/twa/tw_cl_intr.c

757 lines
21 KiB
C
Raw Normal View History

/*
* Copyright (c) 2004-07 Applied Micro Circuits Corporation.
* Copyright (c) 2004-05 Vinod Kashyap
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. 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.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR 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 AUTHOR 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.
*
* $FreeBSD$
*/
/*
* AMCC'S 3ware driver for 9000 series storage controllers.
*
* Author: Vinod Kashyap
* Modifications by: Adam Radford
* Modifications by: Manjunath Ranganathaiah
*/
/*
* Common Layer interrupt handling functions.
*/
#include "tw_osl_share.h"
#include "tw_cl_share.h"
#include "tw_cl_fwif.h"
#include "tw_cl_ioctl.h"
#include "tw_cl.h"
#include "tw_cl_externs.h"
#include "tw_osl_ioctl.h"
/*
* Function name: twa_interrupt
* Description: Interrupt handler. Determines the kind of interrupt,
* and returns TW_CL_TRUE if it recognizes the interrupt.
*
* Input: ctlr_handle -- controller handle
* Output: None
* Return value: TW_CL_TRUE -- interrupt recognized
* TW_CL_FALSE-- interrupt not recognized
*/
TW_INT32
tw_cl_interrupt(struct tw_cl_ctlr_handle *ctlr_handle)
{
struct tw_cli_ctlr_context *ctlr =
(struct tw_cli_ctlr_context *)(ctlr_handle->cl_ctlr_ctxt);
TW_UINT32 status_reg;
TW_INT32 rc = TW_CL_FALSE;
tw_cli_dbg_printf(10, ctlr_handle, tw_osl_cur_func(), "entered");
/* If we don't have controller context, bail */
if (ctlr == NULL)
goto out;
/*
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
* Bail If we get an interrupt while resetting, or shutting down.
*/
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
if (ctlr->reset_in_progress || !(ctlr->active))
goto out;
/* Read the status register to determine the type of interrupt. */
status_reg = TW_CLI_READ_STATUS_REGISTER(ctlr_handle);
if (tw_cli_check_ctlr_state(ctlr, status_reg))
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
goto out;
/* Clear the interrupt. */
if (status_reg & TWA_STATUS_HOST_INTERRUPT) {
tw_cli_dbg_printf(6, ctlr_handle, tw_osl_cur_func(),
"Host interrupt");
TW_CLI_WRITE_CONTROL_REGISTER(ctlr_handle,
TWA_CONTROL_CLEAR_HOST_INTERRUPT);
}
if (status_reg & TWA_STATUS_ATTENTION_INTERRUPT) {
tw_cli_dbg_printf(6, ctlr_handle, tw_osl_cur_func(),
"Attention interrupt");
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
rc |= TW_CL_TRUE; /* request for a deferred isr call */
tw_cli_process_attn_intr(ctlr);
TW_CLI_WRITE_CONTROL_REGISTER(ctlr_handle,
TWA_CONTROL_CLEAR_ATTENTION_INTERRUPT);
}
if (status_reg & TWA_STATUS_COMMAND_INTERRUPT) {
tw_cli_dbg_printf(6, ctlr_handle, tw_osl_cur_func(),
"Command interrupt");
rc |= TW_CL_TRUE; /* request for a deferred isr call */
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
tw_cli_process_cmd_intr(ctlr);
if ((TW_CL_Q_FIRST_ITEM(&(ctlr->req_q_head[TW_CLI_PENDING_Q]))) == TW_CL_NULL)
TW_CLI_WRITE_CONTROL_REGISTER(ctlr_handle,
TWA_CONTROL_MASK_COMMAND_INTERRUPT);
}
if (status_reg & TWA_STATUS_RESPONSE_INTERRUPT) {
tw_cli_dbg_printf(10, ctlr_handle, tw_osl_cur_func(),
"Response interrupt");
rc |= TW_CL_TRUE; /* request for a deferred isr call */
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
tw_cli_process_resp_intr(ctlr);
}
out:
return(rc);
}
/*
* Function name: tw_cli_process_host_intr
* Description: This function gets called if we triggered an interrupt.
* We don't use it as of now.
*
* Input: ctlr -- ptr to CL internal ctlr context
* Output: None
* Return value: None
*/
TW_VOID
tw_cli_process_host_intr(struct tw_cli_ctlr_context *ctlr)
{
tw_cli_dbg_printf(6, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
}
/*
* Function name: tw_cli_process_attn_intr
* Description: This function gets called if the fw posted an AEN
* (Asynchronous Event Notification). It fetches
* all the AEN's that the fw might have posted.
*
* Input: ctlr -- ptr to CL internal ctlr context
* Output: None
* Return value: None
*/
TW_VOID
tw_cli_process_attn_intr(struct tw_cli_ctlr_context *ctlr)
{
TW_INT32 error;
tw_cli_dbg_printf(6, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
if ((error = tw_cli_get_aen(ctlr))) {
/*
* If the driver is already in the process of retrieveing AEN's,
* we will be returned TW_OSL_EBUSY. In this case,
* tw_cli_param_callback or tw_cli_aen_callback will eventually
* retrieve the AEN this attention interrupt is for. So, we
* don't need to print the failure.
*/
if (error != TW_OSL_EBUSY)
tw_cl_create_event(ctlr->ctlr_handle, TW_CL_FALSE,
TW_CL_MESSAGE_SOURCE_COMMON_LAYER_ERROR,
0x1200, 0x1, TW_CL_SEVERITY_ERROR_STRING,
"Failed to fetch AEN",
"error = %d", error);
}
}
/*
* Function name: tw_cli_process_cmd_intr
* Description: This function gets called if we hit a queue full
* condition earlier, and the fw is now ready for
* new cmds. Submits any pending requests.
*
* Input: ctlr -- ptr to CL internal ctlr context
* Output: None
* Return value: None
*/
TW_VOID
tw_cli_process_cmd_intr(struct tw_cli_ctlr_context *ctlr)
{
tw_cli_dbg_printf(6, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
/* Start any requests that might be in the pending queue. */
tw_cli_submit_pending_queue(ctlr);
/*
* If tw_cli_submit_pending_queue was unsuccessful due to a "cmd queue
* full" condition, cmd_intr will already have been unmasked by
* tw_cli_submit_cmd. We don't need to do it again... simply return.
*/
}
/*
* Function name: tw_cli_process_resp_intr
* Description: Looks for cmd completions from fw; queues cmds completed
* by fw into complete queue.
*
* Input: ctlr -- ptr to CL internal ctlr context
* Output: None
* Return value: 0 -- no ctlr error
* non-zero-- ctlr error
*/
TW_INT32
tw_cli_process_resp_intr(struct tw_cli_ctlr_context *ctlr)
{
TW_UINT32 resp;
struct tw_cli_req_context *req;
TW_INT32 error;
TW_UINT32 status_reg;
tw_cli_dbg_printf(10, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
for (;;) {
status_reg = TW_CLI_READ_STATUS_REGISTER(ctlr->ctlr_handle);
if ((error = tw_cli_check_ctlr_state(ctlr, status_reg)))
break;
if (status_reg & TWA_STATUS_RESPONSE_QUEUE_EMPTY) {
tw_cli_dbg_printf(7, ctlr->ctlr_handle,
tw_osl_cur_func(), "Response queue empty");
break;
}
/* Response queue is not empty. */
resp = TW_CLI_READ_RESPONSE_QUEUE(ctlr->ctlr_handle);
{
req = &(ctlr->req_ctxt_buf[GET_RESP_ID(resp)]);
}
if (req->state != TW_CLI_REQ_STATE_BUSY) {
tw_cl_create_event(ctlr->ctlr_handle, TW_CL_FALSE,
TW_CL_MESSAGE_SOURCE_COMMON_LAYER_ERROR,
0x1201, 0x1, TW_CL_SEVERITY_ERROR_STRING,
"Unposted command completed!!",
"request = %p, status = %d",
req, req->state);
#ifdef TW_OSL_DEBUG
tw_cl_print_ctlr_stats(ctlr->ctlr_handle);
#endif /* TW_OSL_DEBUG */
continue;
}
/*
* Remove the request from the busy queue, mark it as complete,
* and enqueue it in the complete queue.
*/
tw_cli_req_q_remove_item(req, TW_CLI_BUSY_Q);
req->state = TW_CLI_REQ_STATE_COMPLETE;
tw_cli_req_q_insert_tail(req, TW_CLI_COMPLETE_Q);
}
/* Complete this, and other requests in the complete queue. */
tw_cli_process_complete_queue(ctlr);
return(error);
}
/*
* Function name: tw_cli_submit_pending_queue
* Description: Kick starts any requests in the pending queue.
*
* Input: ctlr -- ptr to CL internal ctlr context
* Output: None
* Return value: 0 -- all pending requests submitted successfully
* non-zero-- otherwise
*/
TW_INT32
tw_cli_submit_pending_queue(struct tw_cli_ctlr_context *ctlr)
{
struct tw_cli_req_context *req;
TW_INT32 error = TW_OSL_ESUCCESS;
tw_cli_dbg_printf(3, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
/*
* Pull requests off the pending queue, and submit them.
*/
while ((req = tw_cli_req_q_remove_head(ctlr, TW_CLI_PENDING_Q)) !=
TW_CL_NULL) {
if ((error = tw_cli_submit_cmd(req))) {
if (error == TW_OSL_EBUSY) {
tw_cli_dbg_printf(2, ctlr->ctlr_handle,
tw_osl_cur_func(),
"Requeueing pending request");
req->state = TW_CLI_REQ_STATE_PENDING;
/*
* Queue the request at the head of the pending
* queue, and break away, so we don't try to
* submit any more requests.
*/
tw_cli_req_q_insert_head(req, TW_CLI_PENDING_Q);
break;
} else {
tw_cl_create_event(ctlr->ctlr_handle,
TW_CL_FALSE,
TW_CL_MESSAGE_SOURCE_COMMON_LAYER_ERROR,
0x1202, 0x1,
TW_CL_SEVERITY_ERROR_STRING,
"Could not start request "
"in pending queue",
"request = %p, opcode = 0x%x, "
"error = %d", req,
GET_OPCODE(req->cmd_pkt->
command.cmd_pkt_9k.res__opcode),
error);
/*
* Set the appropriate error and call the CL
* internal callback if there's one. If the
* request originator is polling for completion,
* he should be checking req->error to
* determine that the request did not go
* through. The request originators are
* responsible for the clean-up.
*/
req->error_code = error;
req->state = TW_CLI_REQ_STATE_COMPLETE;
if (req->tw_cli_callback)
req->tw_cli_callback(req);
error = TW_OSL_ESUCCESS;
}
}
}
return(error);
}
/*
* Function name: tw_cli_process_complete_queue
* Description: Calls the CL internal callback routine, if any, for
* each request in the complete queue.
*
* Input: ctlr -- ptr to CL internal ctlr context
* Output: None
* Return value: None
*/
TW_VOID
tw_cli_process_complete_queue(struct tw_cli_ctlr_context *ctlr)
{
struct tw_cli_req_context *req;
tw_cli_dbg_printf(10, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
/*
* Pull commands off the completed list, dispatch them appropriately.
*/
while ((req = tw_cli_req_q_remove_head(ctlr, TW_CLI_COMPLETE_Q)) !=
TW_CL_NULL) {
/* Call the CL internal callback, if there's one. */
if (req->tw_cli_callback)
req->tw_cli_callback(req);
}
}
/*
* Function name: tw_cli_complete_io
* Description: CL internal callback for SCSI/fw passthru requests.
*
* Input: req -- ptr to CL internal request context
* Output: None
* Return value: None
*/
TW_VOID
tw_cli_complete_io(struct tw_cli_req_context *req)
{
struct tw_cli_ctlr_context *ctlr = req->ctlr;
struct tw_cl_req_packet *req_pkt =
(struct tw_cl_req_packet *)(req->orig_req);
tw_cli_dbg_printf(8, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
req_pkt->status = TW_CL_ERR_REQ_SUCCESS;
if (req->error_code) {
req_pkt->status = TW_CL_ERR_REQ_UNABLE_TO_SUBMIT_COMMAND;
goto out;
}
if (req->state != TW_CLI_REQ_STATE_COMPLETE) {
tw_cl_create_event(ctlr->ctlr_handle, TW_CL_FALSE,
TW_CL_MESSAGE_SOURCE_COMMON_LAYER_ERROR,
0x1203, 0x1, TW_CL_SEVERITY_ERROR_STRING,
"I/O completion on incomplete command!!",
"request = %p, status = %d",
req, req->state);
#ifdef TW_OSL_DEBUG
tw_cl_print_ctlr_stats(ctlr->ctlr_handle);
#endif /* TW_OSL_DEBUG */
return;
}
if (req->flags & TW_CLI_REQ_FLAGS_PASSTHRU) {
/* Copy the command packet back into OSL's space. */
tw_osl_memcpy(req_pkt->gen_req_pkt.pt_req.cmd_pkt, req->cmd_pkt,
sizeof(struct tw_cl_command_packet));
} else
tw_cli_scsi_complete(req);
out:
req_pkt->tw_osl_callback(req->req_handle);
tw_cli_req_q_insert_tail(req, TW_CLI_FREE_Q);
}
/*
* Function name: tw_cli_scsi_complete
* Description: Completion routine for SCSI requests.
*
* Input: req -- ptr to CL internal request context
* Output: None
* Return value: None
*/
TW_VOID
tw_cli_scsi_complete(struct tw_cli_req_context *req)
{
struct tw_cl_req_packet *req_pkt =
(struct tw_cl_req_packet *)(req->orig_req);
struct tw_cl_scsi_req_packet *scsi_req =
&(req_pkt->gen_req_pkt.scsi_req);
struct tw_cl_command_9k *cmd =
&(req->cmd_pkt->command.cmd_pkt_9k);
struct tw_cl_command_header *cmd_hdr;
TW_UINT16 error;
TW_UINT8 *cdb;
tw_cli_dbg_printf(8, req->ctlr->ctlr_handle, tw_osl_cur_func(),
"entered");
scsi_req->scsi_status = cmd->status;
if (! cmd->status)
return;
tw_cli_dbg_printf(1, req->ctlr->ctlr_handle, tw_osl_cur_func(),
"req_id = 0x%x, status = 0x%x",
GET_REQ_ID(cmd->lun_l4__req_id), cmd->status);
cmd_hdr = &(req->cmd_pkt->cmd_hdr);
error = cmd_hdr->status_block.error;
if ((error == TWA_ERROR_LOGICAL_UNIT_NOT_SUPPORTED) ||
(error == TWA_ERROR_UNIT_OFFLINE)) {
if (GET_LUN_L4(cmd->lun_l4__req_id))
req_pkt->status |= TW_CL_ERR_REQ_INVALID_LUN;
else
req_pkt->status |= TW_CL_ERR_REQ_INVALID_TARGET;
} else {
tw_cli_dbg_printf(2, req->ctlr->ctlr_handle,
tw_osl_cur_func(),
"cmd = %x %x %x %x %x %x %x",
GET_OPCODE(cmd->res__opcode),
GET_SGL_OFF(cmd->res__opcode),
cmd->unit,
cmd->lun_l4__req_id,
cmd->status,
cmd->sgl_offset,
cmd->lun_h4__sgl_entries);
cdb = (TW_UINT8 *)(cmd->cdb);
tw_cli_dbg_printf(2, req->ctlr->ctlr_handle,
tw_osl_cur_func(),
"cdb = %x %x %x %x %x %x %x %x "
"%x %x %x %x %x %x %x %x",
cdb[0], cdb[1], cdb[2], cdb[3],
cdb[4], cdb[5], cdb[6], cdb[7],
cdb[8], cdb[9], cdb[10], cdb[11],
cdb[12], cdb[13], cdb[14], cdb[15]);
#if 0
/*
* Print the error. Firmware doesn't yet support
* the 'Mode Sense' cmd. Don't print if the cmd
* is 'Mode Sense', and the error is 'Invalid field
* in CDB'.
*/
if (! ((cdb[0] == 0x1A) && (error == 0x10D)))
tw_cli_create_ctlr_event(req->ctlr,
TW_CL_MESSAGE_SOURCE_CONTROLLER_ERROR,
cmd_hdr);
#endif // 0
}
if (scsi_req->sense_data) {
tw_osl_memcpy(scsi_req->sense_data, cmd_hdr->sense_data,
TWA_SENSE_DATA_LENGTH);
scsi_req->sense_len = TWA_SENSE_DATA_LENGTH;
req_pkt->status |= TW_CL_ERR_REQ_AUTO_SENSE_VALID;
}
req_pkt->status |= TW_CL_ERR_REQ_SCSI_ERROR;
}
/*
* Function name: tw_cli_param_callback
* Description: Callback for get/set_param requests.
*
* Input: req -- ptr to completed request pkt
* Output: None
* Return value: None
*/
TW_VOID
tw_cli_param_callback(struct tw_cli_req_context *req)
{
struct tw_cli_ctlr_context *ctlr = req->ctlr;
union tw_cl_command_7k *cmd =
&(req->cmd_pkt->command.cmd_pkt_7k);
TW_INT32 error;
tw_cli_dbg_printf(4, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
/*
* If the request was never submitted to the controller, the function
* that sets req->error is responsible for calling tw_cl_create_event.
*/
if (! req->error_code)
if (cmd->param.status) {
#if 0
tw_cli_create_ctlr_event(ctlr,
TW_CL_MESSAGE_SOURCE_CONTROLLER_ERROR,
&(req->cmd_pkt->cmd_hdr));
#endif // 0
tw_cl_create_event(ctlr->ctlr_handle, TW_CL_FALSE,
TW_CL_MESSAGE_SOURCE_COMMON_LAYER_ERROR,
0x1204, 0x1, TW_CL_SEVERITY_ERROR_STRING,
"get/set_param failed",
"status = %d", cmd->param.status);
}
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
ctlr->internal_req_busy = TW_CL_FALSE;
tw_cli_req_q_insert_tail(req, TW_CLI_FREE_Q);
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
if ((ctlr->get_more_aens) && (!(ctlr->reset_in_progress))) {
ctlr->get_more_aens = TW_CL_FALSE;
tw_cli_dbg_printf(4, ctlr->ctlr_handle, tw_osl_cur_func(),
"Fetching more AEN's");
if ((error = tw_cli_get_aen(ctlr)))
tw_cl_create_event(ctlr->ctlr_handle, TW_CL_FALSE,
TW_CL_MESSAGE_SOURCE_COMMON_LAYER_ERROR,
0x1205, 0x1, TW_CL_SEVERITY_ERROR_STRING,
"Failed to fetch all AEN's from param_callback",
"error = %d", error);
}
}
/*
* Function name: tw_cli_aen_callback
* Description: Callback for requests to fetch AEN's.
*
* Input: req -- ptr to completed request pkt
* Output: None
* Return value: None
*/
TW_VOID
tw_cli_aen_callback(struct tw_cli_req_context *req)
{
struct tw_cli_ctlr_context *ctlr = req->ctlr;
struct tw_cl_command_header *cmd_hdr;
struct tw_cl_command_9k *cmd =
&(req->cmd_pkt->command.cmd_pkt_9k);
TW_UINT16 aen_code = TWA_AEN_QUEUE_EMPTY;
TW_INT32 error;
tw_cli_dbg_printf(4, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
tw_cli_dbg_printf(4, ctlr->ctlr_handle, tw_osl_cur_func(),
"req_id = 0x%x, req error = %d, status = 0x%x",
GET_REQ_ID(cmd->lun_l4__req_id), req->error_code, cmd->status);
/*
* If the request was never submitted to the controller, the function
* that sets error is responsible for calling tw_cl_create_event.
*/
if (!(error = req->error_code))
if ((error = cmd->status)) {
cmd_hdr = (struct tw_cl_command_header *)
(&(req->cmd_pkt->cmd_hdr));
#if 0
tw_cli_create_ctlr_event(ctlr,
TW_CL_MESSAGE_SOURCE_CONTROLLER_ERROR,
cmd_hdr);
#endif // 0
tw_cl_create_event(ctlr->ctlr_handle, TW_CL_FALSE,
TW_CL_MESSAGE_SOURCE_COMMON_LAYER_ERROR,
0x1206, 0x1, TW_CL_SEVERITY_ERROR_STRING,
"Request Sense failed",
"opcode = 0x%x, status = %d",
GET_OPCODE(cmd->res__opcode), cmd->status);
}
if (error) {
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
ctlr->internal_req_busy = TW_CL_FALSE;
tw_cli_req_q_insert_tail(req, TW_CLI_FREE_Q);
return;
}
tw_cli_dbg_printf(4, ctlr->ctlr_handle, tw_osl_cur_func(),
"Request Sense command succeeded");
aen_code = tw_cli_manage_aen(ctlr, req);
if (aen_code != TWA_AEN_SYNC_TIME_WITH_HOST) {
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
ctlr->internal_req_busy = TW_CL_FALSE;
tw_cli_req_q_insert_tail(req, TW_CLI_FREE_Q);
if (aen_code != TWA_AEN_QUEUE_EMPTY)
if ((error = tw_cli_get_aen(ctlr)))
tw_cl_create_event(ctlr->ctlr_handle,
TW_CL_FALSE,
TW_CL_MESSAGE_SOURCE_COMMON_LAYER_ERROR,
0x1207, 0x1,
TW_CL_SEVERITY_ERROR_STRING,
"Failed to fetch all AEN's",
"error = %d", error);
}
}
/*
* Function name: tw_cli_manage_aen
* Description: Handles AEN's.
*
* Input: ctlr -- ptr to CL internal ctlr context
* req -- ptr to CL internal request context
* Output: None
* Return value: None
*/
TW_UINT16
tw_cli_manage_aen(struct tw_cli_ctlr_context *ctlr,
struct tw_cli_req_context *req)
{
struct tw_cl_command_header *cmd_hdr;
TW_UINT16 aen_code;
TW_TIME local_time;
TW_TIME sync_time;
TW_UINT32 error;
tw_cli_dbg_printf(4, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
cmd_hdr = (struct tw_cl_command_header *)(req->data);
aen_code = cmd_hdr->status_block.error;
switch (aen_code) {
case TWA_AEN_SYNC_TIME_WITH_HOST:
tw_cli_dbg_printf(4, ctlr->ctlr_handle, tw_osl_cur_func(),
"Received AEN_SYNC_TIME");
/*
* Free the internal req pkt right here, since
* tw_cli_set_param will need it.
*/
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
ctlr->internal_req_busy = TW_CL_FALSE;
tw_cli_req_q_insert_tail(req, TW_CLI_FREE_Q);
/*
* We will use a callback in tw_cli_set_param only when
* interrupts are enabled and we can expect our callback
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
* to get called. Setting the get_more_aens
* flag will make the callback continue to try to retrieve
* more AEN's.
*/
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
if (ctlr->interrupts_enabled)
ctlr->get_more_aens = TW_CL_TRUE;
/* Calculate time (in seconds) since last Sunday 12.00 AM. */
local_time = tw_osl_get_local_time();
sync_time = (local_time - (3 * 86400)) % 604800;
if ((error = tw_cli_set_param(ctlr, TWA_PARAM_TIME_TABLE,
TWA_PARAM_TIME_SCHED_TIME, 4,
&sync_time,
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
(ctlr->interrupts_enabled)
? tw_cli_param_callback : TW_CL_NULL)))
tw_cl_create_event(ctlr->ctlr_handle, TW_CL_FALSE,
TW_CL_MESSAGE_SOURCE_COMMON_LAYER_ERROR,
0x1208, 0x1, TW_CL_SEVERITY_ERROR_STRING,
"Unable to sync time with ctlr",
"error = %d", error);
break;
case TWA_AEN_QUEUE_EMPTY:
tw_cli_dbg_printf(4, ctlr->ctlr_handle, tw_osl_cur_func(),
"AEN queue empty");
break;
default:
/* Queue the event. */
tw_cli_dbg_printf(4, ctlr->ctlr_handle, tw_osl_cur_func(),
"Queueing AEN");
tw_cli_create_ctlr_event(ctlr,
TW_CL_MESSAGE_SOURCE_CONTROLLER_EVENT,
cmd_hdr);
break;
} /* switch */
return(aen_code);
}
/*
* Function name: tw_cli_enable_interrupts
* Description: Enables interrupts on the controller
*
* Input: ctlr -- ptr to CL internal ctlr context
* Output: None
* Return value: None
*/
TW_VOID
tw_cli_enable_interrupts(struct tw_cli_ctlr_context *ctlr)
{
tw_cli_dbg_printf(3, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
ctlr->interrupts_enabled = TW_CL_TRUE;
TW_CLI_WRITE_CONTROL_REGISTER(ctlr->ctlr_handle,
TWA_CONTROL_CLEAR_ATTENTION_INTERRUPT |
TWA_CONTROL_UNMASK_RESPONSE_INTERRUPT |
TWA_CONTROL_ENABLE_INTERRUPTS);
}
/*
* Function name: twa_setup
* Description: Disables interrupts on the controller
*
* Input: ctlr -- ptr to CL internal ctlr context
* Output: None
* Return value: None
*/
TW_VOID
tw_cli_disable_interrupts(struct tw_cli_ctlr_context *ctlr)
{
tw_cli_dbg_printf(3, ctlr->ctlr_handle, tw_osl_cur_func(), "entered");
TW_CLI_WRITE_CONTROL_REGISTER(ctlr->ctlr_handle,
TWA_CONTROL_DISABLE_INTERRUPTS);
Apply driver update from LSI. Many thanks to LSI for continuing to support FreeBSD. 1) Timeout ioctl command timeouts. Do not reset the controller if ioctl command completed successfully. 2) Remove G66_WORKAROUND code (this bug never shipped). 3) Remove unnecessary interrupt lock (intr_lock). 4) Timeout firmware handshake for PChip reset (don't wait forever). 5) Handle interrupts inline. 6) Unmask command interrupt ONLY when adding a command to the pending queue. 7) Mask command interrupt ONLY after removing the last command from the pending queue. 8) Remove TW_OSLI_DEFERRED_INTR_USED code. 9) Replace controller "state" with separate data fields to avoid races: TW_CLI_CTLR_STATE_ACTIVE ctlr->active TW_CLI_CTLR_STATE_INTR_ENABLED ctlr->interrupts_enabled TW_CLI_CTLR_STATE_INTERNAL_REQ_BUSY ctlr->internal_req_busy TW_CLI_CTLR_STATE_GET_MORE_AENS ctlr->get_more_aens TW_CLI_CTLR_STATE_RESET_IN_PROGRESS ctlr->reset_in_progress TW_CLI_CTLR_STATE_RESET_PHASE1_IN_PROGRESS ctlr->reset_phase1_in_progress 10) Fix "req" leak in twa_action() when simq is frozen and req is NOT null. 11) Replace softc "state" with separate data fields to avoid races: TW_OSLI_CTLR_STATE_OPEN sc->open TW_OSLI_CTLR_STATE_SIMQ_FROZEN sc->simq_frozen 12) Fix reference to TW_OSLI_REQ_FLAGS_IN_PROGRESS in tw_osl_complete_passthru() 13) Use correct CAM status values. Change CAM_REQ_CMP_ERR to CAM_REQ_INVALID. Remove use of CAM_RELEASE_SIMQ for physical data addresses. 14) Do not freeze/ release the simq with non I/O commands. When it is appropriate to temporarily freeze the simq with an I/O command use: xpt_freeze_simq(sim, 1); ccb->ccb_h.status |= CAM_RELEASE_SIMQ; otherwise use: xpt_freeze_simq(sim, 1); xpt_release_simq(sim, 1); Submitted by: Tom Couch <tom.couch lsi.com> PR: kern/147695 MFC after: 3 days
2010-06-09 21:40:38 +00:00
ctlr->interrupts_enabled = TW_CL_FALSE;
}