[PATCH 05/10] ipr: Disk remove path cleanup

[PATCH 05/10] ipr: Disk remove path cleanup

[Brian King]

Instead of NULLing the resource entry pointer when a disk
goes away to prevent any new commands being sent to it,
set the adapter resource handle to an invalid value so
new ops getting sent to it will fail with a selection timeout
response. This patch is needed for future SATA patches.


Signed-off-by: Brian King <brking@us.ibm.com>
---

 libata-dev-bjking1/drivers/scsi/ipr.c |    6 +++---
 libata-dev-bjking1/drivers/scsi/ipr.h |    1 +
 2 files changed, 4 insertions(+), 3 deletions(-)

diff -puN drivers/scsi/ipr.c~ipr_disk_remove_cleanup drivers/scsi/ipr.c
--- libata-dev/drivers/scsi/ipr.c~ipr_disk_remove_cleanup	2006-03-06 09:20:06.000000000 -0600
+++ libata-dev-bjking1/drivers/scsi/ipr.c	2006-03-06 09:20:06.000000000 -0600
@@ -869,8 +869,8 @@ static void ipr_handle_config_change(str
 
 	if (hostrcb->hcam.notify_type == IPR_HOST_RCB_NOTIF_TYPE_REM_ENTRY) {
 		if (res->sdev) {
-			res->sdev->hostdata = NULL;
 			res->del_from_ml = 1;
+			res->cfgte.res_handle = IPR_INVALID_RES_HANDLE;
 			if (ioa_cfg->allow_ml_add_del)
 				schedule_work(&ioa_cfg->work_q);
 		} else
@@ -2107,7 +2107,6 @@ restart:
 				did_work = 1;
 				sdev = res->sdev;
 				if (!scsi_device_get(sdev)) {
-					res->sdev = NULL;
 					list_move_tail(&res->queue, &ioa_cfg->free_res_q);
 					spin_unlock_irqrestore(ioa_cfg->host->host_lock, lock_flags);
 					scsi_remove_device(sdev);
@@ -2124,6 +2123,7 @@ restart:
 			bus = res->cfgte.res_addr.bus;
 			target = res->cfgte.res_addr.target;
 			lun = res->cfgte.res_addr.lun;
+			res->add_to_ml = 0;
 			spin_unlock_irqrestore(ioa_cfg->host->host_lock, lock_flags);
 			scsi_add_device(ioa_cfg->host, bus, target, lun);
 			spin_lock_irqsave(ioa_cfg->host->host_lock, lock_flags);
@@ -4980,7 +4980,7 @@ static int ipr_init_res_table(struct ipr
 	list_for_each_entry_safe(res, temp, &old_res, queue) {
 		if (res->sdev) {
 			res->del_from_ml = 1;
-			res->sdev->hostdata = NULL;
+			res->cfgte.res_handle = IPR_INVALID_RES_HANDLE;
 			list_move_tail(&res->queue, &ioa_cfg->used_res_q);
 		} else {
 			list_move_tail(&res->queue, &ioa_cfg->free_res_q);
diff -puN drivers/scsi/ipr.h~ipr_disk_remove_cleanup drivers/scsi/ipr.h
--- libata-dev/drivers/scsi/ipr.h~ipr_disk_remove_cleanup	2006-03-06 09:20:06.000000000 -0600
+++ libata-dev-bjking1/drivers/scsi/ipr.h	2006-03-06 09:20:06.000000000 -0600
@@ -133,6 +133,7 @@
 #define IPR_MAX_SCSI_RATE(width) ((320 * 10) / ((width) / 8))
 
 #define IPR_IOA_RES_HANDLE				0xffffffff
+#define IPR_INVALID_RES_HANDLE			0
 #define IPR_IOA_RES_ADDR				0x00ffffff
 
 /*
_

Re: [PATCH 05/10] ipr: Disk remove path cleanup

[Christoph Hellwig]

On Wed, Mar 29, 2006 at 08:38:18AM -0600, Brian King wrote:
> 
> Instead of NULLing the resource entry pointer when a disk
> goes away to prevent any new commands being sent to it,
> set the adapter resource handle to an invalid value so
> new ops getting sent to it will fail with a selection timeout
> response. This patch is needed for future SATA patches.

It would help greatly if you could split out the patches that don't
require libata changes so they could go into scsi-misc already.

Re: [PATCH 05/10] ipr: Disk remove path cleanup

[Brian King]

Christoph Hellwig wrote:
> On Wed, Mar 29, 2006 at 08:38:18AM -0600, Brian King wrote:
>> Instead of NULLing the resource entry pointer when a disk
>> goes away to prevent any new commands being sent to it,
>> set the adapter resource handle to an invalid value so
>> new ops getting sent to it will fail with a selection timeout
>> response. This patch is needed for future SATA patches.
> 
> It would help greatly if you could split out the patches that don't
> require libata changes so they could go into scsi-misc already.

Ok. Will do.

-- 
Brian King
eServer Storage I/O
IBM Linux Technology Center