Messages in this thread |  | | Subject | Re: [4.15-rc9] fs_reclaim lockdep trace | From | Tetsuo Handa <> | Date | Sun, 28 Jan 2018 13:25:29 +0900 |
| |
On 2018/01/28 10:16, Tetsuo Handa wrote: > Linus Torvalds wrote: >> On Sat, Jan 27, 2018 at 2:24 PM, Dave Jones <davej@codemonkey.org.uk> wrote: >>> On Tue, Jan 23, 2018 at 08:36:51PM -0500, Dave Jones wrote: >>> > Just triggered this on a server I was rsync'ing to. >>> >>> Actually, I can trigger this really easily, even with an rsync from one >>> disk to another. Though that also smells a little like networking in >>> the traces. Maybe netdev has ideas. >> >> Is this new to 4.15? Or is it just that you're testing something new? >> >> If it's new and easy to repro, can you just bisect it? And if it isn't >> new, can you perhaps check whether it's new to 4.14 (ie 4.13 being >> ok)? >> >> Because that fs_reclaim_acquire/release() debugging isn't new to 4.15, >> but it was rewritten for 4.14.. I'm wondering if that remodeling ended >> up triggering something. > > --- linux-4.13.16/mm/page_alloc.c > +++ linux-4.14.15/mm/page_alloc.c
Oops. This output was inverted.
> @@ -3527,53 +3519,12 @@ > return true; > } > return false; > } > #endif /* CONFIG_COMPACTION */ > > -#ifdef CONFIG_LOCKDEP > -struct lockdep_map __fs_reclaim_map = > - STATIC_LOCKDEP_MAP_INIT("fs_reclaim", &__fs_reclaim_map); > - > -static bool __need_fs_reclaim(gfp_t gfp_mask) > -{ > - gfp_mask = current_gfp_context(gfp_mask); > - > - /* no reclaim without waiting on it */ > - if (!(gfp_mask & __GFP_DIRECT_RECLAIM)) > - return false; > - > - /* this guy won't enter reclaim */ > - if ((current->flags & PF_MEMALLOC) && !(gfp_mask & __GFP_NOMEMALLOC)) > - return false;
Since __kmalloc_reserve() from __alloc_skb() adds __GFP_NOMEMALLOC | __GFP_NOWARN to gfp_mask, __need_fs_reclaim() is failing to return false here.
But why checking __GFP_NOMEMALLOC here? __alloc_pages_slowpath() skips direct reclaim if !(gfp_mask & __GFP_DIRECT_RECLAIM) or (current->flags & PF_MEMALLOC), doesn't it?
---------- static inline struct page * __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order, struct alloc_context *ac) { (...snipped...) /* Caller is not willing to reclaim, we can't balance anything */ if (!can_direct_reclaim) goto nopage;
/* Avoid recursion of direct reclaim */ if (current->flags & PF_MEMALLOC) goto nopage;
/* Try direct reclaim and then allocating */ page = __alloc_pages_direct_reclaim(gfp_mask, order, alloc_flags, ac, &did_some_progress); if (page) goto got_pg; (...snipped...) } ----------
|  |