From: Gabriel Krisman Bertazi
Date: Wed, 27 May 2015 13:51:43 -0300
> Subject: [PATCH] bnx2x: Alloc 4k fragment for each rx ring buffer element
>
> The driver allocates one page for each buffer on the rx ring, which is
> too much on architectures like ppc64 and can cause unexpecte
imple linear relationship between the two.
Yes, but I felt it would make the code more readable, even though it
introduced sobre redundancy. Anyway, the new version below removes
frag_count and fixes the overflow bug.
Thanks for your review!
-- >8 --
Subject: [PATCH] bnx2x: Alloc 4k fragment for
Hi Gabriel,
Thanks for pointing out this shortcoming.
However, I don't think that the solution you suggested would work as is.
@@ -628,21 +644,30 @@ static int bnx2x_fill_frag_skb(struct bnx2x *bp, struct
bnx2x_fastpath *fp,
bnx2x_fp_qstats(bp, fp)->rx_skb_alloc_failed++;
The driver allocates one page for each buffer on the rx ring, which is
too much on architectures like ppc64 and can cause unexpected allocation
failures when the system is under stress. Now, if the architecture's
PAGE_SIZE is greater than 4k, we fragment pages and assign each 4k
segment to a ring
