Messages in this thread | | | Date | Thu, 13 Mar 2014 15:59:42 -0600 | From | Jens Axboe <> | Subject | Re: [PATCH] mtip32xx: mtip_async_complete() bug fixes |
| |
On 03/13/2014 03:33 PM, Sam Bradshaw wrote: > On 03/13/2014 02:02 PM, Jens Axboe wrote: >> On 03/13/2014 03:09 PM, Sam Bradshaw wrote: >>> This patch fixes 2 issues in the fast completion path: >>> 1) Possible double completions / double dma_unmap_sg() calls due to lack >>> of atomicity in the check and subsequent dereference of the upper layer >>> callback function. Fixed with cmpxchg before unmap and callback. >>> 2) Regression in unaligned IO constraining workaround for p420m devices. >>> Fixed by checking if IO is unaligned and using proper semaphore if so. >> >> Sam, what is this patch against? It doesn't apply to -rc6 and it doesn't >> apply to for-3.15/drivers (the latter would be preferred/required). It >> also seems to have line break issues. >> > > It's against for-3.15/drivers. Sorry about the breaks; if I remove them > does it apply cleanly?
Yep, this one is much better. Thanks!
-- Jens Axboe
| |