Re: [GIT] Networking

[Geert Uytterhoeven]

On Fri, Jun 13, 2014 at 8:59 AM, Johannes Berg
<johannes@xxxxxxxxxxxxxxxx> wrote:
> On Fri, 2014-06-13 at 06:48 +0000, Linus Torvalds wrote:
>> On Thu, Jun 12, 2014 at 12:14 PM, David Miller <davem@xxxxxxxxxxxxx> wrote:
>> >
>> >   git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git master
>>
>> Hmm. I get *lots* of the appended messages from iwlwifi now. Things
>> still seem to work, but ...
>>
>> This is a Haswell laptop with bog-standard intel wireless:
>>
>>   01:00.0 Network controller: Intel Corporation Wireless 7260 (rev 6b)
>>
>> Any ideas?
>
> Looks odd.
>
> If the stacktrace/message
>
>> [  273.783438] <unknown>: hw csum failure
>> [  273.783449] CPU: 2 PID: 2199 Comm: Chrome_IOThread Not tainted 3.15.0-11906-g6d87c225f5d8 #8
>> [  273.783451] Hardware name: Sony Corporation SVP11213CXB/VAIO, BIOS R0270V7 05/17/2013
>> [  273.783461] Call Trace:
>> [  273.783477]  [<ffffffff8159d199>] ? dump_stack+0x41/0x51
>> [  273.783488]  [<ffffffff814a4f1d>] ? skb_copy_and_csum_datagram_iovec+0xfd/0x110
>> [  273.783494]  [<ffffffff8150e2ea>] ? udp_recvmsg+0x1ca/0x390
>> [  273.783500]  [<ffffffff81518e95>] ? inet_recvmsg+0x65/0x80
>> [  273.783506]  [<ffffffff814971f5>] ? sock_recvmsg+0x95/0xd0
>> [  273.783511]  [<ffffffff81197862>] ? ep_send_events_proc+0xa2/0x1b0
>> [  273.783513]  [<ffffffff811977c0>] ? ep_read_events_proc+0xb0/0xb0
>> [  273.783516]  [<ffffffff81497325>] ? SYSC_recvfrom+0xb5/0x110
>> [  273.783521]  [<ffffffff8108bdc0>] ? wake_up_state+0x10/0x10
>> [  273.783527]  [<ffffffff815a3d92>] ? system_call_fastpath+0x16/0x1b
>
> is to be believed, then I'm not sure why you think iwlwifi is involved
> though.
>
> Either way, iwlwifi doesn't support HW checksumming, so it failing makes
> no sense, it's not there to start with.
>
> The only thing that touched this seems to be Tom's patch:
>
> commit 7e3cead5172927732f51fde77fef6f521e22f209
> Author: Tom Herbert <therbert@xxxxxxxxxx>
> Date:   Tue Jun 10 18:54:19 2014 -0700
>
>     net: Save software checksum complete
>
> but I'm not familiar enough to say where this leads.
>
> If it's iwlwifi, then maybe that commit broke something with devices
> that don't support checksumming?

Tom posted some fixes for this:
http://marc.info/?l=linux-netdev&m=140261417832399&w=2

I can confirm the first one fixes the issue I was having (10s delay in
ssh login), cfr. http://marc.info/?l=linux-netdev&m=140257250715322&w=2

However, the second patch reintroduces the bad behavior again?

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/