+struct bnx2x_alloc_pool {
+ struct page *page;
+ dma_addr_t dma;
+ u8 offset;
+ u8 frag_count;
+};
...
static int bnx2x_alloc_rx_sge(struct bnx2x *bp, struct bnx2x_fastpath *fp,
u16 index, gfp_t gfp_mask)
{
Yuval Mintz yuval.mi...@qlogic.com writes:
Actually, this upsets me greatly. We didn't see it on a system with 4KB
pages, but this means you've actually tried to 'sell' us a fastpath fix that
was never tested on machines for which it was meant as an improvement.
The iteration that inserted
On 05/21/2015 03:20 PM, Gabriel Krisman Bertazi wrote:
+#define SGE_PAGE_SHIFT 12
+#define SGE_PAGE_SIZE(1 SGE_PAGE_SHIFT)
...
+struct bnx2x_alloc_pool {
+ struct page *page;
+ dma_addr_t dma;
+ u8 offset;
+ u8
Subject: Re: [PATCH v4] bnx2x: Alloc 4k fragment for each rx ring buffer
element
Yuval Mintz yuval.mi...@qlogic.com writes:
Regardless, I'll give it a more thorough review tomorrow.
[If those are all the problems we'll find with it, I don't think
we'll need to re-spin this once more
On 21.05.2015 15:20, Gabriel Krisman Bertazi wrote:
Yuval Mintz yuval.mi...@qlogic.com writes:
Regardless, I'll give it a more thorough review tomorrow.
[If those are all the problems we'll find with it, I don't think we'll
need to re-spin this once more; that is, unless Dave insists]
As
issues that Yuval
pointed out. Are there any other concerns regarding this one?
-- 8 --
Subject: [PATCH v4] 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 unexpected