| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
The return type of block_device_operations->release() changed from
int to void in commit db2a144bedd58b3dcf19950c2f476c58c9f39d18
Fixes the following compiler warning:
sbull/sbull.c:313:2: warning:
initialization from incompatible pointer type [enabled by default]
sbull/sbull.c:313:2: warning:
(near initialization for ‘sbull_ops.release’) [enabled by default]
|
|
|
|
|
| |
sbull/sbull.c:194:2: warning:
‘return’ with a value, in function returning void [enabled by default]
|
|
|
|
| |
Signed-off-by: Javier Martinez Canillas <javier@dowhile0.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The block layer request API changes make a requirement for block drivers
to dequeue requests before signal its completition.
Commit 1dfd4ab138fc9f9cad869e6110022c7cfd5544d1 changed sbull to use
the new block layer API. But the patch only renamed the functions while
keeping the old driver semantics. It used blk_peek_request() that does the
same that elv_next_request() did.
Remove the request from the queue before is not only a recomendation
but a true requirement. Attempting to complete a request that remains
in the queue make the kernel oops. So the last commit introduced a Bug.
This patch uses blk_fetch_request() instead of blk_peek_request().
blk_fetch_request() not only returns the request in the head of the
queue but also removes the request from the queue so a latter
signaling of the request completition doesn't hang the system.
|
|
|
|
|
| |
This wrapper doesn't exist anymore in current kernel (2.6.37) so
so doesn't make any sense to know its existance.
|
|
|
|
| |
kobject_put(q->kobj) instead
|
| |
|
|
|
|
| |
macro __rq_for_each_bio()
|
| |
|
| |
|
|
|
|
| |
(http://lwn.net/Articles/333620/)
|
| |
|
| |
|
| |
|
|
|