... protect against > this, but rc1's [SCSI] use dynamically allocated sense buffer > ... I tried a few things to fix it. > Adding a no-op constructor to the sense_slab inhibits ... is to revert the separate sense_slab, using > cache-line-aligned sense_buffer ...

linux-kernel - James Bottomley - Apr 6 2008 - 19:35

... is to revert the separate sense_slab, using > cache-line-aligned sense_buffer ... bios piled up above a scsi host working one by > one ... fairly soon with the livelock fix. > So lacking better ideas, stick ... and James agreed, the patch in scsi-misc looks good to ...

linux-kernel - FUJITA Tomonori - Apr 6 2008 - 22:48

... the __scsi_get_command end, there are GFP_ATOMIC sense_slab > allocations, which do give access to ... allocated from __GFP_HIGH > reserves to a scsi sense_slab, swap_writepages are liable to ... line of my swap over network patches I made 'robustified' SLAB to ...

linux-kernel - Peter Zijlstra - Apr 7 2008 - 15:55

... to merge > > both bio cache and sense_slab cache into kmalloc's 128-byte ... the __scsi_get_command end, there are GFP_ATOMIC sense_slab allocations, which do give access to ... from __GFP_HIGH reserves to a scsi sense_slab, swap_writepages are liable to gobble ...

linux-kernel - Hugh Dickins - Apr 7 2008 - 15:40

... to protect against > this, but rc1's [SCSI] use dynamically allocated sense buffer > ... it fails to allocate from the separate sense_slab, instead > of giving up, it must fall ... happens to merge > both bio cache and sense_slab cache into kmalloc's 128-byte ...

linux-kernel - Christoph Lameter - Apr 7 2008 - 01:26

... alternative is to revert the separate sense_slab, using > > cache-line-aligned sense_buffer allocated ... the known problem. > > Reverting the separate sense_slab is fine for now but we ... the one big buffer, and you're better off with the separate slabs: ...

linux-kernel - Hugh Dickins - Apr 7 2008 - 14:07

... alternative is to revert the separate sense_slab, using > > > cache-line-aligned sense_buffer allocated ... the known problem. > > > > Reverting the separate sense_slab is fine for now but we ... the one big > buffer, and you're better off with the separate ...

linux-kernel - FUJITA Tomonori - Apr 8 2008 - 10:04

... allocated from __GFP_HIGH > > reserves to a scsi sense_slab, swap_writepages are liable to gobble ... line of my swap over network patches I made > 'robustified' SLAB to ensure ... has been dealt with (small fix to scsi free_list handling: which resembles a ...

linux-kernel - Hugh Dickins - Apr 7 2008 - 16:31

... been allocated from __GFP_HIGH > > > reserves to a scsi sense_slab, swap_writepages are liable to gobble up ... the line of my swap over network patches I made > > 'robustified' SLAB to ensure these ... and it's generally accepted we're in a slow path once we ...

linux-kernel - Peter Zijlstra - Apr 7 2008 - 16:47

... 11:16:33 2008 -0500 > > > > > > [SCSI] consolidate command allocation in a ... 25-rc8 (plus cfq rcu > > fix). Not quite what you asked ... to -rc8 plus cfq rcu fix; but I'm not entirely ... effectively reversing your patch, so testing current scsi-misc is very ...

linux-kernel - James Bottomley - Apr 7 2008 - 14:04

... was sort of accidentally fixed in scsi-misc by commit > > commit ... 11:16:33 2008 -0500 > > [SCSI] consolidate command allocation in a ... pieces to backport. Precisely that patch seems appropriate to 2.6 ... to -rc8 plus cfq rcu fix; but I'm not entirely ...

linux-kernel - Hugh Dickins - Apr 6 2008 - 21:01

... > This was sort of accidentally fixed in scsi-misc by commit > > > > commit ... Mar 13 11:16:33 2008 -0500 > > > > [SCSI] consolidate command allocation in a single place ... > pieces to backport. > > Precisely that patch seems appropriate to 2.6.25- ...

linux-kernel - Hugh Dickins - Apr 7 2008 - 13:51

... > > Somewhat irrelevant since Linus put my patch into 2.6.25-rc8-git > > ( ... > Well ... just remember that to merge scsi-misc with what Linus has now > done, I'm effectively reversing your patch, so testing current scsi-misc > is very valuable if you ...

linux-kernel - Hugh Dickins - Apr 7 2008 - 14:26

... it has been dealt with (small fix to scsi > free_list handling: which resembles a mempool, but done its own way) ... and why we don't need to fix it for the SLOB case as well ... term. I would prefer some other fix even if it means getting rid of ...

linux-kernel - Pekka Enberg - Apr 7 2008 - 17:00

... Enberg wrote: > So something like this fugly patch on top of the SLUB variable order > patches that let the allocator fall-back to smaller page orders. ... exist with the page allocator and a fix there would fix it for all users. -- To ...

linux-kernel - Christoph Lameter - Apr 8 2008 - 16:45

... with the page allocator and a fix > there would fix it for all users. This ... anyway. Keep in mind that the patch is specifically addressing writeback in OOM conditions ... bringing up a good point that my patch doesn't actually _guarantee_ anything so I ...

linux-kernel - Pekka Enberg - Apr 8 2008 - 17:11

... > acceptable in an embedded environment and why we don't need to fix it > for the SLOB case as well. > > But you do ... > measure for the short term. I would prefer some other fix even if it > means getting rid of slab merging competely (which would ...

linux-kernel - Pekka Enberg - Apr 7 2008 - 17:05

... why we don't need to fix it for the SLOB case as ... then of course it should be fixed in SLOB also. I just meant ... . I would prefer some other fix even if it means getting > ... stronger reasons than that if we're to pursue this. Hugh -- To unsubscribe ...

linux-kernel - Hugh Dickins - Apr 7 2008 - 17:30

... page allocator and a fix > > there would fix it for all ... Keep in mind that > the patch is specifically addressing writeback in ... a good point that my patch doesn't > actually _guarantee_ anything ... allocation as long as we're in the emergency state. Also ...

linux-kernel - Peter Zijlstra - Apr 8 2008 - 17:40

... my logs: I'd better find stronger reasons than that if we're to > pursue this. No, I do agree that it's a problem. I just don't like the proposed no-merge fix ;-). Pekka -- To unsubscribe from this list: send the line "unsubscribe linux- ...

linux-kernel - Pekka Enberg - Apr 7 2008 - 17:36