linux-kernel - Re: [PATCH net-next 2/3] net: use skb_for_each

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 9 Apr 2021 22:44:50 +0200
From:   Matteo Croce <mcroce@...ux.microsoft.com>
To:     Jakub Kicinski <kuba@...nel.org>
Cc:     netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
        "David S. Miller" <davem@...emloft.net>,
        Julia Lawall <julia.lawall@...ia.fr>
Subject: Re: [PATCH net-next 2/3] net: use skb_for_each_frag() helper where possible

On Fri, Apr 9, 2021 at 8:54 PM Jakub Kicinski <kuba@...nel.org> wrote:
>
> On Fri,  9 Apr 2021 20:06:04 +0200 Matteo Croce wrote:
> > From: Matteo Croce <mcroce@...rosoft.com>
> >
> > use the new helper macro skb_for_each_frag() which allows to iterate
> > through all the SKB fragments.
> >
> > The patch was created with Coccinelle, this was the semantic patch:
>
> Bunch of set but not used warnings here. Please make sure the code
> builds cleanly allmodconfig, W=1 C=1 before posting.
>

Will do.

> What pops to mind (although quite nit picky) is the question if the
> assembly changes much between driver which used to cache nr_frags and
> now always going skb_shinfo(skb)->nr_frags? It's a relatively common
> pattern.

Since skb_shinfo() is a macro and skb_end_pointer() a static inline,
it should be the same, but I was curious to check so, this is a diff
between the following snippet before and afer the macro:

int frags = skb_shinfo(skb)->nr_frags;
int i;
for (i = 0; i < frags; i++)
    kfree(skb->frags[i]);

 1 file changed, 8 insertions(+), 7 deletions(-)

--- ins1.s 2021-04-09 22:35:59.384523865 +0200
+++ ins2.s 2021-04-09 22:36:08.132594737 +0200
@@ -1,26 +1,27 @@
 iter:
         movsx   rax, DWORD PTR [rdi+16]
         mov     rdx, QWORD PTR [rdi+8]
         mov     eax, DWORD PTR [rdx+rax]
         test    eax, eax
         jle     .L6
         push    rbp
-        sub     eax, 1
+        mov     rbp, rdi
         push    rbx
-        lea     rbp, [rdi+32+rax*8]
-        lea     rbx, [rdi+24]
+        xor     ebx, ebx
         sub     rsp, 8
 .L3:
-        mov     rdi, QWORD PTR [rbx]
-        add     rbx, 8
+        mov     rdi, QWORD PTR [rbp+24+rbx*8]
+        add     rbx, 1
         call    kfree
-        cmp     rbx, rbp
-        jne     .L3
+        movsx   rax, DWORD PTR [rbp+16]
+        mov     rdx, QWORD PTR [rbp+8]
+        cmp     DWORD PTR [rdx+rax], ebx
+        jg      .L3
         add     rsp, 8
         xor     eax, eax
         pop     rbx
         pop     rbp
         ret
 .L6:
         xor     eax, eax
         ret

-- 
per aspera ad upstream