Re: 2.6.14-rc5 e1000 and page allocation failures.. still

[Robert Hancock <hancockr@shaw.ca>]

John Bäckstrand wrote:
> Im seeing a massive amount of page allocation failures with 2.6.14-rc5, 
> and also earlier kernels, see "E1000 - page allocation failure - saga 
> continues :(". Machine is a 1Ghz Athlon with 256MB RAM. Attached is 
> example dmesg output. The stack traces come in many variants. Killing 
> processes using RAM only seems to help temporarily. Ive also tried 
> setting vm.min_free_kbytes=16384, which used to work pretty well, but 
> this does not help (atleast not in the state the machine is currently 
> in, without rebooting).
>
> free currently gives:
>
>              total       used       free     shared    buffers     cached
> Mem:        256520     239128      17392          0       5372      67500
> -/+ buffers/cache:     166256      90264
> Swap:       506036      21248     484788
>
>
> I havent yet tried rebooting and using the vm.min_free_kbytes=16384 from 
> scratch, but I think something with the default for this is wrong if it 
> results in this many page allocation errors. The machine is serving 
> files from an encrypted partition with reiserfs on it, and I obivously 
> use the e1000 driver.
>
> ---
> John Bäckstrand
>
>
> ------------------------------------------------------------------------
>
> [149649.847890] kcryptd/0: page allocation failure. order:3, mode:0x20

..

> [149649.849933] Free pages:       16148kB (0kB HighMem)

It looks like you have enough memory free - the problem is that the 
driver is allocating a block of memory with order 3, which is 8 pages. 
Quite likely there are not enough contiguous free pages to satisfy that.

That's an awful big buffer size for a packet - I assume you're using 
jumbo frames or something? Ideally the driver and hardware should be 
able to allocate a buffer for those packets in multiple chunks, but I 
have no idea if this is possible.

--
Robert Hancock      Saskatoon, SK, Canada
To email, remove "nospam" from hancockr@nospamshaw.ca
Home Page: http://www.roberthancock.com/

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/