Christian König
2021-Feb-10 10:52 UTC
[Nouveau] drm/nouneau: 5.11 cycle regression bisected to 461619f5c324 "drm/nouveau: switch to new allocator"
Am 10.02.21 um 11:46 schrieb Mike Galbraith:> On Wed, 2021-02-10 at 11:34 +0100, Christian K?nig wrote: >> What seems to happen here is that your system is low on resources and we >> just try to free up pages. > FWIW, box has oodles generic ram free right after boot.Then I have no idea what happens here. The returned pages are identical, only the overhead to allocate them is reduced significantly. The only other idea I have is that this somehow changes the timing and exposes a bug somewhere else. You could try to replace the "for (order = min(MAX_ORDER - 1UL, __fls(num_pages)); num_pages;" in ttm_pool_alloc() with "for (order = 0; num_pages;" to get the old behavior. Regards, Christian.> > -Mike >
Mike Galbraith
2021-Feb-10 11:22 UTC
[Nouveau] drm/nouneau: 5.11 cycle regression bisected to 461619f5c324 "drm/nouveau: switch to new allocator"
On Wed, 2021-02-10 at 11:52 +0100, Christian K?nig wrote:> > > You could try to replace the "for (order = min(MAX_ORDER - 1UL, > __fls(num_pages)); num_pages;" in ttm_pool_alloc() with "for (order = 0; > num_pages;" to get the old behavior.That's a nogo too. -Mike