[PATCH] cciss: Add missing allocation in scsi_cmd_stack_setup and corresponding deallocation

[PATCH] cciss: Add missing allocation in scsi_cmd_stack_setup and corresponding deallocation

[Stephen M. Cameron]

From: Stephen M. Cameron <scameron@beardog.cce.hp.com>

This bit got lost somewhere along the way.  Without this, panic.

Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
---
 drivers/block/cciss_scsi.c |    9 ++++++++-
 1 files changed, 8 insertions(+), 1 deletions(-)

diff --git a/drivers/block/cciss_scsi.c b/drivers/block/cciss_scsi.c
index 514470d..6961002 100644
--- a/drivers/block/cciss_scsi.c
+++ b/drivers/block/cciss_scsi.c
@@ -222,7 +222,12 @@ scsi_cmd_stack_setup(ctlr_info_t *h, struct cciss_scsi_adapter_data_t *sa)
 		sa->cmd_sg_list = NULL;
 		return -ENOMEM;
 	}
-
+	stk->elem = kmalloc(sizeof(stk->elem[0]) * stk->nelems, GFP_KERNEL);
+	if (!stk->elem) {
+		pci_free_consistent(h->pdev, size, stk->pool,
+		stk->cmd_pool_handle);
+		return -1;
+	}
 	for (i = 0; i < stk->nelems; i++) {
 		stk->elem[i] = &stk->pool[i];
 		stk->elem[i]->busaddr = (__u32) (stk->cmd_pool_handle + 
@@ -252,6 +257,8 @@ scsi_cmd_stack_free(ctlr_info_t *h)
 	pci_free_consistent(h->pdev, size, stk->pool, stk->cmd_pool_handle);
 	stk->pool = NULL;
 	cciss_free_sg_chain_blocks(sa->cmd_sg_list, stk->nelems);
+	kfree(stk->elem);
+	stk->elem = NULL;
 }
 
 #if 0

Re: [PATCH] cciss: Add missing allocation in scsi_cmd_stack_setup and corresponding deallocation

[Jens Axboe]

On 2011-03-10 21:22, Stephen M. Cameron wrote:
> From: Stephen M. Cameron <scameron@beardog.cce.hp.com>
> 
> This bit got lost somewhere along the way.  Without this, panic.

This doesn't apply cleanly, neither hunk. I hand applied it. Please let
me know what a patch is against in the future, if it doesn't apply
"anywhere".

I marked this one as stable backport, too.

-- 
Jens Axboe

Re: [PATCH] cciss: Add missing allocation in scsi_cmd_stack_setup and corresponding deallocation

[scameron]

On Fri, Mar 11, 2011 at 08:09:39PM +0100, Jens Axboe wrote:
> On 2011-03-10 21:22, Stephen M. Cameron wrote:
> > From: Stephen M. Cameron <scameron@beardog.cce.hp.com>
> > 
> > This bit got lost somewhere along the way.  Without this, panic.
> 
> This doesn't apply cleanly, neither hunk. I hand applied it. Please let
> me know what a patch is against in the future, if it doesn't apply
> "anywhere".

Oh... I see what happened.

That patch was fixing something which was accidentally 
left out of this patch:

http://marc.info/?l=linux-kernel&m=129840572232151&w=2

which I mistakenly thought was present in 2.6.38-rc7,
but it's not present (quite likely because without the
missing bit, it causes a panic), It's sitting in my stack of
patches still.

[scameron@localhost linux-2.6.38-rc7]$ stg series | grep cciss
+ cciss-hoist-tag-masking-out-of-loop
+ cciss-inform-controller-we-are-using-32-bit-tags
+ cciss-mask-off-error-bits-before-calling-pci-free-consistent
+ cciss-remove-unnecessary-casts
+ cciss-fix-missed-command-status-value-CMD_UNABORTABLE
+ cciss-add-cciss_tape_cmds-module-parameter
+ cciss-fix-missing-initialization-in-scsi-cmd-stack-setup
+ cciss-export-resettable-host-attribute
+ cciss-fix-sense-data-copying-code

What needs to happen is I combine cciss-add-cciss_tape_cmds-module-parameter
and cciss-fix-missing-initialization-in-scsi-cmd-stack-setup into one 
working patch.

Sorry about that.  I need to slow down.  I'm making too many mistakes.

-- steve



> 
> I marked this one as stable backport, too.
> 
> -- 
> Jens Axboe

Re: [PATCH] cciss: Add missing allocation in scsi_cmd_stack_setup and corresponding deallocation

[Jens Axboe]

On 2011-03-11 20:46, scameron@beardog.cce.hp.com wrote:
> On Fri, Mar 11, 2011 at 08:09:39PM +0100, Jens Axboe wrote:
>> On 2011-03-10 21:22, Stephen M. Cameron wrote:
>>> From: Stephen M. Cameron <scameron@beardog.cce.hp.com>
>>>
>>> This bit got lost somewhere along the way.  Without this, panic.
>>
>> This doesn't apply cleanly, neither hunk. I hand applied it. Please let
>> me know what a patch is against in the future, if it doesn't apply
>> "anywhere".
> 
> Oh... I see what happened.
> 
> That patch was fixing something which was accidentally 
> left out of this patch:
> 
> http://marc.info/?l=linux-kernel&m=129840572232151&w=2
> 
> which I mistakenly thought was present in 2.6.38-rc7,
> but it's not present (quite likely because without the
> missing bit, it causes a panic), It's sitting in my stack of
> patches still.
> 
> [scameron@localhost linux-2.6.38-rc7]$ stg series | grep cciss
> + cciss-hoist-tag-masking-out-of-loop
> + cciss-inform-controller-we-are-using-32-bit-tags
> + cciss-mask-off-error-bits-before-calling-pci-free-consistent
> + cciss-remove-unnecessary-casts
> + cciss-fix-missed-command-status-value-CMD_UNABORTABLE
> + cciss-add-cciss_tape_cmds-module-parameter
> + cciss-fix-missing-initialization-in-scsi-cmd-stack-setup
> + cciss-export-resettable-host-attribute
> + cciss-fix-sense-data-copying-code
> 
> What needs to happen is I combine cciss-add-cciss_tape_cmds-module-parameter
> and cciss-fix-missing-initialization-in-scsi-cmd-stack-setup into one 
> working patch.

Please check what is missing for my for-2.6.39/drivers branch and send
me the patches.

-- 
Jens Axboe