Messages in this thread | | | Date | Mon, 3 Jul 2000 02:57:39 +0200 | From | Jens Axboe <> | Subject | Re: [PATCH] 2.2.17pre9 interactiveness under high IO |
| |
On Sun, Jul 02 2000, Marcelo Tosatti wrote: > > This is wrong -- if the buffer is not queued because we are out of > > free request slots, it has still already been locked at this point > > (and _Req). > > if (!req) { > if (rw_ahead) > goto end_io; > req = __get_request_wait(max_req, bh->b_rdev); > } > > In case we fail to find a free request, bh->b_end_io is called, and > it unlocks the buffer.
Ah yes, I was comparing to a similar case in 2.4. The buffer is still marked BH_Req though, and thus counted as busy. Have you thoroughly tested this change in 2.2 and made sure it didn't cause any ext2 corruption? I saw some scary stuff in 2.4
-- * Jens Axboe <axboe@suse.de> * SuSE Labs
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |