On Thu, 1 Jun 2006 08:05:49 +0300 (EEST)
Kai Makisara <[email protected]> wrote:
> On Wed, 31 May 2006, James Bottomley wrote:
>
> > This is my current slew of small bug fixes which either fix serious
> > bugs, or are completely safe for this -rc5 stage of the kernel. I've
> > added one more since I last sent you this pull request (the fix memory
> > building non-aligned sg lists)
> >
> > The patch is available from:
> >
> > master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6.git
> >
> > The short changelog is:
> >
> > Bryan Holty:
> > o fix memory building non-aligned sg lists
> >
> I looked at
> www.kernel.org/git/?p=linux/kernel/git/jejb/scsi-rc-fixes-2.6.git;.
>
> This patch does the following change:
>
> - int nr_pages = (bufflen + PAGE_SIZE - 1) >> PAGE_SHIFT;
> + int nr_pages = PAGE_ALIGN(bufflen + sgl[0].offset);
>
> This seems to wrong: the new version is missing the right shift. For
> instance, offset=0 and bufflen=4096 results in 4096 and not 1!
>
> (Using asm-x86_64, the new version translates to
> ((bufflen + sgl[0].offset+PAGE_SIZE-1)&(~(PAGE_SIZE-1)))
> )
>
> According to the original patch by Brian, the change should probably have
> been to (or something equivalent):
>
> + int nr_pages = (bufflen + sgl[0].offset + PAGE_SIZE - 1) >>
> PAGE_SHIFT;
>
> This was tested by several people. Did anyone test the version put into
> scsi-rc-fixes-2.6.git?
>
argh, that was me "improving" things.
--- devel/drivers/scsi/scsi_lib.c~scsi-properly-count-the-number-of-pages-in-scsi_req_map_sg-fix 2006-05-31 22:22:12.000000000 -0700
+++ devel-akpm/drivers/scsi/scsi_lib.c 2006-05-31 22:22:34.000000000 -0700
@@ -367,7 +367,7 @@ static int scsi_req_map_sg(struct reques
int nsegs, unsigned bufflen, gfp_t gfp)
{
struct request_queue *q = rq->q;
- int nr_pages = PAGE_ALIGN(bufflen + sgl[0].offset);
+ int nr_pages = (bufflen + sgl[0].offset + PAGE_SIZE - 1) >> PAGE_SHIFT;
unsigned int data_len = 0, len, bytes, off;
struct page *page;
struct bio *bio = NULL;
_
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]
