* [PATCH] - mptfusion - adding back the spin locks in eh handlers
@ 2005-06-23 18:20 Moore, Eric Dean
2005-06-23 19:04 ` Luben Tuikov
0 siblings, 1 reply; 2+ messages in thread
From: Moore, Eric Dean @ 2005-06-23 18:20 UTC (permalink / raw)
To: linux scsi; +Cc: James Bottomley
The eh handlers are called with host_lock acquired, and interrupts disabled.
However fusion waits for completion of task managment request. With these spin locks
removed, the driver is going to hang within the sleep calls. Someone recently posted
a patch to remove them(don´t know who), and can be found in James Bottomely scsi-misc branch.
Please apply.
Signed-off-by: Eric Moore <Eric.Moore@lsil.com>
diff -uarN b/drivers/message/fusion/mptscsih.c a/drivers/message/fusion/mptscsih.c
--- b/drivers/message/fusion/mptscsih.c 2005-06-23 11:18:44.000000000 -0600
+++ a/drivers/message/fusion/mptscsih.c 2005-05-10 14:27:13.000000000 -0600
@@ -1707,6 +1707,7 @@
MPT_FRAME_HDR *mf;
u32 ctx2abort;
int scpnt_idx;
+ spinlock_t *host_lock = SCpnt->device->host->host_lock;
/* If we can't locate our host adapter structure, return FAILED status.
*/
@@ -1754,6 +1755,7 @@
hd->abortSCpnt = SCpnt;
+ spin_unlock_irq(host_lock);
if (mptscsih_TMHandler(hd, MPI_SCSITASKMGMT_TASKTYPE_ABORT_TASK,
SCpnt->device->channel, SCpnt->device->id, SCpnt->device->lun,
ctx2abort, 2 /* 2 second timeout */)
@@ -1770,6 +1772,8 @@
hd->tmPending = 0;
hd->tmState = TM_STATE_NONE;
+ spin_lock_irq(host_lock);
+
/* Unmap the DMA buffers, if any. */
if (SCpnt->use_sg) {
pci_unmap_sg(ioc->pcidev, (struct scatterlist *) SCpnt->request_buffer,
@@ -1785,6 +1789,7 @@
mpt_free_msg_frame(ioc, mf);
return FAILED;
}
+ spin_lock_irq(host_lock);
return SUCCESS;
}
@@ -1801,6 +1806,7 @@
mptscsih_dev_reset(struct scsi_cmnd * SCpnt)
{
MPT_SCSI_HOST *hd;
+ spinlock_t *host_lock = SCpnt->device->host->host_lock;
/* If we can't locate our host adapter structure, return FAILED status.
*/
@@ -1817,6 +1823,7 @@
printk(KERN_WARNING MYNAM ": %s: >> Attempting target reset! (sc=%p)\n",
hd->ioc->name, SCpnt);
+ spin_unlock_irq(host_lock);
if (mptscsih_TMHandler(hd, MPI_SCSITASKMGMT_TASKTYPE_TARGET_RESET,
SCpnt->device->channel, SCpnt->device->id,
0, 0, 5 /* 5 second timeout */)
@@ -1828,10 +1835,12 @@
hd->ioc->name, SCpnt);
hd->tmPending = 0;
hd->tmState = TM_STATE_NONE;
+ spin_lock_irq(host_lock);
return FAILED;
}
-
+ spin_lock_irq(host_lock);
return SUCCESS;
+
}
/*=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=*/
@@ -1865,6 +1874,7 @@
hd->timeouts++;
/* We are now ready to execute the task management request. */
+ spin_unlock_irq(host_lock);
if (mptscsih_TMHandler(hd, MPI_SCSITASKMGMT_TASKTYPE_RESET_BUS,
SCpnt->device->channel, 0, 0, 0, 5 /* 5 second timeout */)
< 0){
@@ -1880,7 +1890,7 @@
spin_lock_irq(host_lock);
return FAILED;
}
-
+ spin_lock_irq(host_lock);
return SUCCESS;
}
@@ -1899,6 +1909,7 @@
{
MPT_SCSI_HOST * hd;
int status = SUCCESS;
+ spinlock_t *host_lock = SCpnt->device->host->host_lock;
/* If we can't locate the host to reset, then we failed. */
if ((hd = (MPT_SCSI_HOST *) SCpnt->device->host->hostdata) == NULL){
@@ -1914,6 +1925,7 @@
/* If our attempts to reset the host failed, then return a failed
* status. The host will be taken off line by the SCSI mid-layer.
*/
+ spin_unlock_irq(host_lock);
if (mpt_HardResetHandler(hd->ioc, CAN_SLEEP) < 0){
status = FAILED;
} else {
@@ -1923,6 +1935,8 @@
hd->tmPending = 0;
hd->tmState = TM_STATE_NONE;
}
+ spin_lock_irq(host_lock);
+
dtmprintk( ( KERN_WARNING MYNAM ": mptscsih_host_reset: "
"Status = %s\n",
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 2+ messages in thread* Re: [PATCH] - mptfusion - adding back the spin locks in eh handlers
2005-06-23 18:20 [PATCH] - mptfusion - adding back the spin locks in eh handlers Moore, Eric Dean
@ 2005-06-23 19:04 ` Luben Tuikov
0 siblings, 0 replies; 2+ messages in thread
From: Luben Tuikov @ 2005-06-23 19:04 UTC (permalink / raw)
To: Eric.Moore; +Cc: linux scsi, James Bottomley
On 06/23/05 14:20, Moore, Eric Dean wrote:
> The eh handlers are called with host_lock acquired, and interrupts disabled.
> However fusion waits for completion of task managment request. With these spin locks
> removed, the driver is going to hang within the sleep calls. Someone recently posted
> a patch to remove them(don´t know who), and can be found in James Bottomely scsi-misc branch.
That was Jeff, and I thought those patches made it in already.
They are needed and are a good thing.
Luben
>
> Please apply.
>
> Signed-off-by: Eric Moore <Eric.Moore@lsil.com>
>
>
> diff -uarN b/drivers/message/fusion/mptscsih.c a/drivers/message/fusion/mptscsih.c
> --- b/drivers/message/fusion/mptscsih.c 2005-06-23 11:18:44.000000000 -0600
> +++ a/drivers/message/fusion/mptscsih.c 2005-05-10 14:27:13.000000000 -0600
> @@ -1707,6 +1707,7 @@
> MPT_FRAME_HDR *mf;
> u32 ctx2abort;
> int scpnt_idx;
> + spinlock_t *host_lock = SCpnt->device->host->host_lock;
>
> /* If we can't locate our host adapter structure, return FAILED status.
> */
> @@ -1754,6 +1755,7 @@
>
> hd->abortSCpnt = SCpnt;
>
> + spin_unlock_irq(host_lock);
> if (mptscsih_TMHandler(hd, MPI_SCSITASKMGMT_TASKTYPE_ABORT_TASK,
> SCpnt->device->channel, SCpnt->device->id, SCpnt->device->lun,
> ctx2abort, 2 /* 2 second timeout */)
> @@ -1770,6 +1772,8 @@
> hd->tmPending = 0;
> hd->tmState = TM_STATE_NONE;
>
> + spin_lock_irq(host_lock);
> +
> /* Unmap the DMA buffers, if any. */
> if (SCpnt->use_sg) {
> pci_unmap_sg(ioc->pcidev, (struct scatterlist *) SCpnt->request_buffer,
> @@ -1785,6 +1789,7 @@
> mpt_free_msg_frame(ioc, mf);
> return FAILED;
> }
> + spin_lock_irq(host_lock);
> return SUCCESS;
> }
>
> @@ -1801,6 +1806,7 @@
> mptscsih_dev_reset(struct scsi_cmnd * SCpnt)
> {
> MPT_SCSI_HOST *hd;
> + spinlock_t *host_lock = SCpnt->device->host->host_lock;
>
> /* If we can't locate our host adapter structure, return FAILED status.
> */
> @@ -1817,6 +1823,7 @@
> printk(KERN_WARNING MYNAM ": %s: >> Attempting target reset! (sc=%p)\n",
> hd->ioc->name, SCpnt);
>
> + spin_unlock_irq(host_lock);
> if (mptscsih_TMHandler(hd, MPI_SCSITASKMGMT_TASKTYPE_TARGET_RESET,
> SCpnt->device->channel, SCpnt->device->id,
> 0, 0, 5 /* 5 second timeout */)
> @@ -1828,10 +1835,12 @@
> hd->ioc->name, SCpnt);
> hd->tmPending = 0;
> hd->tmState = TM_STATE_NONE;
> + spin_lock_irq(host_lock);
> return FAILED;
> }
> -
> + spin_lock_irq(host_lock);
> return SUCCESS;
> +
> }
>
> /*=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=*/
> @@ -1865,6 +1874,7 @@
> hd->timeouts++;
>
> /* We are now ready to execute the task management request. */
> + spin_unlock_irq(host_lock);
> if (mptscsih_TMHandler(hd, MPI_SCSITASKMGMT_TASKTYPE_RESET_BUS,
> SCpnt->device->channel, 0, 0, 0, 5 /* 5 second timeout */)
> < 0){
> @@ -1880,7 +1890,7 @@
> spin_lock_irq(host_lock);
> return FAILED;
> }
> -
> + spin_lock_irq(host_lock);
> return SUCCESS;
> }
>
> @@ -1899,6 +1909,7 @@
> {
> MPT_SCSI_HOST * hd;
> int status = SUCCESS;
> + spinlock_t *host_lock = SCpnt->device->host->host_lock;
>
> /* If we can't locate the host to reset, then we failed. */
> if ((hd = (MPT_SCSI_HOST *) SCpnt->device->host->hostdata) == NULL){
> @@ -1914,6 +1925,7 @@
> /* If our attempts to reset the host failed, then return a failed
> * status. The host will be taken off line by the SCSI mid-layer.
> */
> + spin_unlock_irq(host_lock);
> if (mpt_HardResetHandler(hd->ioc, CAN_SLEEP) < 0){
> status = FAILED;
> } else {
> @@ -1923,6 +1935,8 @@
> hd->tmPending = 0;
> hd->tmState = TM_STATE_NONE;
> }
> + spin_lock_irq(host_lock);
> +
>
> dtmprintk( ( KERN_WARNING MYNAM ": mptscsih_host_reset: "
> "Status = %s\n",
> -
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2005-06-23 19:04 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-06-23 18:20 [PATCH] - mptfusion - adding back the spin locks in eh handlers Moore, Eric Dean
2005-06-23 19:04 ` Luben Tuikov
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox