Hi,
Two questions:
1. I notice that in the GIT version of core/ucm.c, the trunk updates below
were not taken into the 2.6.17 GIT. Is there a reason?
------------------------------------------------------------------------
r5643 | sean.hefty | 2006-03-07 02:09:48 +0200 (Tue, 07 Mar 2006) | 4 lines
Changed paths:
M /gen2/trunk/src/linux-kernel/infiniband/core/ucm.c
Convert file mutex from a semaphore to a real mutex for 2.6.16.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
------------------------------------------------------------------------
------------------------------------------------------------------------
r5944 | roland | 2006-03-22 03:02:22 +0200 (Wed, 22 Mar 2006) | 4 lines
Changed paths:
M /gen2/trunk/src/linux-kernel/infiniband/core/ucm.c
file_mutex is really a struct mutex, so use mutex_init() instead of
init_MUTEX().
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
------------------------------------------------------------------------
2. Missing mutex protection in ucm.c for list_empty call. Is there a reason?
GIT:
static unsigned int ib_ucm_poll(struct file *filp,
struct poll_table_struct *wait)
{
struct ib_ucm_file *file = filp->private_data;
unsigned int mask = 0;
poll_wait(filp, &file->poll_wait, wait);
if (!list_empty(&file->events))
mask = POLLIN | POLLRDNORM;
return mask;
}
Trunk:
static unsigned int ib_ucm_poll(struct file *filp,
struct poll_table_struct *wait)
{
struct ib_ucm_file *file = filp->private_data;
unsigned int mask = 0;
poll_wait(filp, &file->poll_wait, wait);
mutex_lock(&file->file_mutex);
if (!list_empty(&file->events))
mask = POLLIN | POLLRDNORM;
mutex_unlock(&file->file_mutex);
return mask;
}
_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
