Make pkcsslotd grab the shared memory it needs, even if it wasnt freed by a previous instance

Make pkcsslotd grab the shared memory it needs, even if it wasnt freed by a previous instance

BUG=1778
TEST=Login, logout, login and verify that pkcsslotd is running

[Will Drewry, 2010-06-02 01:20:39]

On 2010/06/02 01:14:07, cmasone wrote:
> 

Where does the shmem patch live?

[Chris Masone, 2010-06-02 02:05:40]

On 2010/06/02 01:20:39, Will Drewry wrote:
> On 2010/06/02 01:14:07, cmasone wrote:
> > 
> 
> Where does the shmem patch live?

forgot it.  PTAL

[Will Drewry, 2010-06-02 02:35:17]

seems fine to me but a few nits that are maybe off base

http://codereview.chromium.org/2471003/diff/4001/5001
File dev-libs/opencryptoki/files/opencryptoki-2.2.8-steal_shmem.patch (right):

http://codereview.chromium.org/2471003/diff/4001/5001#newcode12
dev-libs/opencryptoki/files/opencryptoki-2.2.8-steal_shmem.patch:12: +     
shmid = shmget( tok, sizeof( Slot_Mgr_Shr_t ), 0);
nit: should there be a trailing space to match the other call below?

http://codereview.chromium.org/2471003/diff/4001/5001#newcode13
dev-libs/opencryptoki/files/opencryptoki-2.2.8-steal_shmem.patch:13: +     
DestroySharedMemory();
Can you still do an ErrLog() so that we know it happened or do you think it'll
be too noisy?

http://codereview.chromium.org/2471003/diff/4001/5001#newcode16
dev-libs/opencryptoki/files/opencryptoki-2.2.8-steal_shmem.patch:16: +	         
    S_IRGRP  | S_IWUSR | S_IWGRP   );
Do you need to recheck the shmid value after this in case of error? (e.g., it
couldn't be destroyed, etc?)

[Chris Masone, 2010-06-02 18:00:38]

http://codereview.chromium.org/2471003/diff/4001/5001
File dev-libs/opencryptoki/files/opencryptoki-2.2.8-steal_shmem.patch (right):

http://codereview.chromium.org/2471003/diff/4001/5001#newcode12
dev-libs/opencryptoki/files/opencryptoki-2.2.8-steal_shmem.patch:12: +     
shmid = shmget( tok, sizeof( Slot_Mgr_Shr_t ), 0);
On 2010/06/02 02:35:18, Will Drewry wrote:
> nit: should there be a trailing space to match the other call below?

Done.

http://codereview.chromium.org/2471003/diff/4001/5001#newcode13
dev-libs/opencryptoki/files/opencryptoki-2.2.8-steal_shmem.patch:13: +     
DestroySharedMemory();
On 2010/06/02 02:35:18, Will Drewry wrote:
> Can you still do an ErrLog() so that we know it happened or do you think it'll
> be too noisy?

Done.

http://codereview.chromium.org/2471003/diff/4001/5001#newcode16
dev-libs/opencryptoki/files/opencryptoki-2.2.8-steal_shmem.patch:16: +	         
    S_IRGRP  | S_IWUSR | S_IWGRP   );
On 2010/06/02 02:35:18, Will Drewry wrote:
> Do you need to recheck the shmid value after this in case of error? (e.g., it
> couldn't be destroyed, etc?)

Done.

[Will Drewry, 2010-06-02 18:37:06]

LGTM with an optional nit

http://codereview.chromium.org/2471003/diff/10002/11001
File dev-libs/opencryptoki/files/opencryptoki-2.2.8-steal_shmem.patch (right):

http://codereview.chromium.org/2471003/diff/10002/11001#newcode33
dev-libs/opencryptoki/files/opencryptoki-2.2.8-steal_shmem.patch:33: -
looks like you dropped an empty line down here.