* r8169 not workin on resume due allocation failure
@ 2010-09-08 3:44 Holger Freyther
2010-09-08 4:11 ` Mitchell Erblich
0 siblings, 1 reply; 2+ messages in thread
From: Holger Freyther @ 2010-09-08 3:44 UTC (permalink / raw)
To: netdev
Hi all,
I'm using the r8169 module on my notebook and after a couple of days
of running with suspend/resume the networking stops working. I have
the dmesg output at the end of the email. From my understanding this
is due that on suspend the interface will be brought down, the rx buffers
are freed and that on resume it is not possible to allocate a continous
buffer for the sizes required.
My approach to handle this would be to add an option to not free the
buffers until the module is unloaded and guard this with an option. Is
that an acceptable solution?
NetworkManager: page allocation failure. order:3, mode:0x4020
Pid: 1291, comm: NetworkManager \n
Not tainted 2.6.35.4-12.fc14.i686.PAE #1
Call Trace:
[<c07e2197>] ? printk+0x25/0x2e
[<c04c1af6>] __alloc_pages_nodemask+0x5bc/0x62f
[<c04c1b83>] __get_free_pages+0x1a/0x2c
[<c04e7bd6>] __kmalloc_track_caller+0x32/0x144
[<c07465fb>] ? __netdev_alloc_skb+0x1b/0x36
[<c07464fd>] __alloc_skb+0x59/0x100
[<c07465fb>] __netdev_alloc_skb+0x1b/0x36
[<f9796a8f>] rtl8169_rx_fill+0x99/0x132 [r8169]
[<f9797913>] rtl8169_init_ring+0x66/0x96 [r8169]
[<f9797e98>] rtl8169_open+0x7f/0x1d7 [r8169]
[<c0750b04>] __dev_open+0x79/0xaa
[<c0750cf9>] __dev_change_flags+0x98/0x10d
[<c0750dde>] dev_change_flags+0x18/0x44
[<c075976f>] do_setlink+0x253/0x52e
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: r8169 not workin on resume due allocation failure
2010-09-08 3:44 r8169 not workin on resume due allocation failure Holger Freyther
@ 2010-09-08 4:11 ` Mitchell Erblich
0 siblings, 0 replies; 2+ messages in thread
From: Mitchell Erblich @ 2010-09-08 4:11 UTC (permalink / raw)
To: Holger Freyther; +Cc: netdev
Holger, et al,
Actually wouldn't it make sense if a system type fix is done?
I would add a delay if not already done and check to see
if some dirty pages could be cleaned, then freed.
The assumption is then those pages would be minor faulted
back in and prefetched.
Another approach is to make sure that pages are cleaned on
a more frequent basis independent on the number of free
pages (free pages may not be continuous) that wants to make
sure higher order allocs are allocatable over time.
Mitchell Erblich
On Sep 7, 2010, at 8:44 PM, Holger Freyther wrote:
> Hi all,
>
> I'm using the r8169 module on my notebook and after a couple of days
> of running with suspend/resume the networking stops working. I have
> the dmesg output at the end of the email. From my understanding this
> is due that on suspend the interface will be brought down, the rx buffers
> are freed and that on resume it is not possible to allocate a continous
> buffer for the sizes required.
>
> My approach to handle this would be to add an option to not free the
> buffers until the module is unloaded and guard this with an option. Is
> that an acceptable solution?
>
>
> NetworkManager: page allocation failure. order:3, mode:0x4020
> Pid: 1291, comm: NetworkManager \n
> Not tainted 2.6.35.4-12.fc14.i686.PAE #1
> Call Trace:
> [<c07e2197>] ? printk+0x25/0x2e
> [<c04c1af6>] __alloc_pages_nodemask+0x5bc/0x62f
> [<c04c1b83>] __get_free_pages+0x1a/0x2c
> [<c04e7bd6>] __kmalloc_track_caller+0x32/0x144
> [<c07465fb>] ? __netdev_alloc_skb+0x1b/0x36
> [<c07464fd>] __alloc_skb+0x59/0x100
> [<c07465fb>] __netdev_alloc_skb+0x1b/0x36
> [<f9796a8f>] rtl8169_rx_fill+0x99/0x132 [r8169]
> [<f9797913>] rtl8169_init_ring+0x66/0x96 [r8169]
> [<f9797e98>] rtl8169_open+0x7f/0x1d7 [r8169]
> [<c0750b04>] __dev_open+0x79/0xaa
> [<c0750cf9>] __dev_change_flags+0x98/0x10d
> [<c0750dde>] dev_change_flags+0x18/0x44
> [<c075976f>] do_setlink+0x253/0x52e
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2010-09-08 4:17 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-09-08 3:44 r8169 not workin on resume due allocation failure Holger Freyther
2010-09-08 4:11 ` Mitchell Erblich
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox