Huge vmalloc higher-order backing pages were allocated with __GFP_COMP
in order to allow the sub-pages to be refcounted by callers such as
"remap_vmalloc_page [sic]" (remap_vmalloc_range).
However a similar problem exists for other struct page fields callers
use, for example fb_deferred_io_fault() takes a vmalloc'ed page and
not only refcounts it but uses ->lru, ->mapping, ->index.
This is not compatible with compound sub-pages, and can cause bad page
state issues like
BUG: Bad page state in process swapper/0 pfn:00743
page:(____ptrval____) refcount:0 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x743
flags: 0x7ffff000000000(node=0|zone=0|lastcpupid=0x7ffff)
raw: 007ffff000000000 c00c00000001d0c8 c00c00000001d0c8 0000000000000000
raw: 0000000000000000 0000000000000000 00000000ffffffff 0000000000000000
page dumped because: corrupted mapping in tail page
Modules linked in:
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.18.0-rc3-00082-gfc6fff4a7ce1-dirty #2810
Call Trace:
dump_stack_lvl+0x74/0xa8 (unreliable)
bad_page+0x12c/0x170
free_tail_pages_check+0xe8/0x190
free_pcp_prepare+0x31c/0x4e0
free_unref_page+0x40/0x1b0
__vunmap+0x1d8/0x420
...
The correct approach is to use split high-order pages for the huge
vmalloc backing. These allow callers to treat them in exactly the same
way as individually-allocated order-0 pages.
Link: https://lore.kernel.org/all/14444103-d51b-0fb3-ee63-c3f182f0b546@molgen.mpg.de/
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Cc: Paul Menzel <pmenzel@molgen.mpg.de>
Cc: Song Liu <songliubraving@fb.com>
Cc: Rick Edgecombe <rick.p.edgecombe@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Huge page backed vmalloc memory could benefit performance in many cases.
However, some users of vmalloc may not be ready to handle huge pages for
various reasons: hardware constraints, potential pages split, etc.
VM_NO_HUGE_VMAP was introduced to allow vmalloc users to opt-out huge
pages. However, it is not easy to track down all the users that require
the opt-out, as the allocation are passed different stacks and may cause
issues in different layers.
To address this issue, replace VM_NO_HUGE_VMAP with an opt-in flag,
VM_ALLOW_HUGE_VMAP, so that users that benefit from huge pages could ask
specificially.
Also, remove vmalloc_no_huge() and add opt-in helper vmalloc_huge().
Fixes: fac54e2bfb ("x86/Kconfig: Select HAVE_ARCH_HUGE_VMALLOC with HAVE_ARCH_HUGE_VMAP")
Link: https://lore.kernel.org/netdev/14444103-d51b-0fb3-ee63-c3f182f0b546@molgen.mpg.de/"
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Song Liu <song@kernel.org>
Reviewed-by: Rik van Riel <riel@surriel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit 3ee48b6af4 ("mm, x86: Saving vmcore with non-lazy freeing of
vmas") introduced set_iounmap_nonlazy(), which sets vmap_lazy_nr to
lazy_max_pages() + 1, ensuring that any future vunmaps() immediately
purge the vmap areas instead of doing it lazily.
Commit 690467c81b ("mm/vmalloc: Move draining areas out of caller
context") moved the purging from the vunmap() caller to a worker thread.
Unfortunately, set_iounmap_nonlazy() can cause the worker thread to spin
(possibly forever). For example, consider the following scenario:
1. Thread reads from /proc/vmcore. This eventually calls
__copy_oldmem_page() -> set_iounmap_nonlazy(), which sets
vmap_lazy_nr to lazy_max_pages() + 1.
2. Then it calls free_vmap_area_noflush() (via iounmap()), which adds 2
pages (one page plus the guard page) to the purge list and
vmap_lazy_nr. vmap_lazy_nr is now lazy_max_pages() + 3, so the
drain_vmap_work is scheduled.
3. Thread returns from the kernel and is scheduled out.
4. Worker thread is scheduled in and calls drain_vmap_area_work(). It
frees the 2 pages on the purge list. vmap_lazy_nr is now
lazy_max_pages() + 1.
5. This is still over the threshold, so it tries to purge areas again,
but doesn't find anything.
6. Repeat 5.
If the system is running with only one CPU (which is typicial for kdump)
and preemption is disabled, then this will never make forward progress:
there aren't any more pages to purge, so it hangs. If there is more
than one CPU or preemption is enabled, then the worker thread will spin
forever in the background. (Note that if there were already pages to be
purged at the time that set_iounmap_nonlazy() was called, this bug is
avoided.)
This can be reproduced with anything that reads from /proc/vmcore
multiple times. E.g., vmcore-dmesg /proc/vmcore.
It turns out that improvements to vmap() over the years have obsoleted
the need for this "optimization". I benchmarked `dd if=/proc/vmcore
of=/dev/null` with 4k and 1M read sizes on a system with a 32GB vmcore.
The test was run on 5.17, 5.18-rc1 with a fix that avoided the hang, and
5.18-rc1 with set_iounmap_nonlazy() removed entirely:
|5.17 |5.18+fix|5.18+removal
4k|40.86s| 40.09s| 26.73s
1M|24.47s| 23.98s| 21.84s
The removal was the fastest (by a wide margin with 4k reads). This
patch removes set_iounmap_nonlazy().
Link: https://lkml.kernel.org/r/52f819991051f9b865e9ce25605509bfdbacadcd.1649277321.git.osandov@fb.com
Fixes: 690467c81b ("mm/vmalloc: Move draining areas out of caller context")
Signed-off-by: Omar Sandoval <osandov@fb.com>
Acked-by: Chris Down <chris@chrisdown.name>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Acked-by: Baoquan He <bhe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add vmalloc tagging support to HW_TAGS KASAN.
The key difference between HW_TAGS and the other two KASAN modes when it
comes to vmalloc: HW_TAGS KASAN can only assign tags to physical memory.
The other two modes have shadow memory covering every mapped virtual
memory region.
Make __kasan_unpoison_vmalloc() for HW_TAGS KASAN:
- Skip non-VM_ALLOC mappings as HW_TAGS KASAN can only tag a single
mapping of normal physical memory; see the comment in the function.
- Generate a random tag, tag the returned pointer and the allocation,
and initialize the allocation at the same time.
- Propagate the tag into the page stucts to allow accesses through
page_address(vmalloc_to_page()).
The rest of vmalloc-related KASAN hooks are not needed:
- The shadow-related ones are fully skipped.
- __kasan_poison_vmalloc() is kept as a no-op with a comment.
Poisoning and zeroing of physical pages that are backing vmalloc()
allocations are skipped via __GFP_SKIP_KASAN_UNPOISON and
__GFP_SKIP_ZERO: __kasan_unpoison_vmalloc() does that instead.
Enabling CONFIG_KASAN_VMALLOC with HW_TAGS is not yet allowed.
Link: https://lkml.kernel.org/r/d19b2e9e59a9abc59d05b72dea8429dcaea739c6.1643047180.git.andreyknvl@google.com
Signed-off-by: Andrey Konovalov <andreyknvl@google.com>
Co-developed-by: Vincenzo Frascino <vincenzo.frascino@arm.com>
Signed-off-by: Vincenzo Frascino <vincenzo.frascino@arm.com>
Acked-by: Marco Elver <elver@google.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Evgenii Stepanov <eugenis@google.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Collingbourne <pcc@google.com>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Extend the find_vmap_lowest_match() function with one more parameter.
It is "adjust_search_size" boolean variable, so it is possible to
control an accuracy of search block if a specific alignment is required.
With this patch, a search size is always adjusted, to serve a request as
fast as possible because of performance reason.
But there is one exception though, it is short ranges where requested
size corresponds to passed vstart/vend restriction together with a
specific alignment request. In such scenario an adjustment wold not
lead to success allocation.
Link: https://lkml.kernel.org/r/20220119143540.601149-2-urezki@gmail.com
Signed-off-by: Uladzislau Rezki <uladzislau.rezki@sony.com>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Oleksiy Avramchenko <oleksiy.avramchenko@sonymobile.com>
Cc: Vasily Averin <vvs@virtuozzo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
A caller initiates the drain procces from its context once the
drain threshold is reached or passed. There are at least two
drawbacks of doing so:
a) a caller can be a high-prio or RT task. In that case it can
stuck in doing the actual drain of all lazily freed areas.
This is not optimal because such tasks usually are latency
sensitive where the control should be returned back as soon
as possible in order to drive such workloads in time. See
96e2db4561 ("mm/vmalloc: rework the drain logic")
b) It is not safe to call vfree() during holding a spinlock due
to the vmap_purge_lock mutex. The was a report about this from
Zeal Robot <zealci@zte.com.cn> here:
https://lore.kernel.org/all/20211222081026.484058-1-chi.minghao@zte.com.cn
Moving the drain to the separate work context addresses those
issues.
v1->v2:
- Added prefix "_work" to the drain worker function.
v2->v3:
- Remove the drain_vmap_work_in_progress. Extra queuing
is expectable under heavy load but it can be disregarded
because a work will bail out if nothing to be done.
Link: https://lkml.kernel.org/r/20220131144058.35608-1-urezki@gmail.com
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Oleksiy Avramchenko <oleksiy.avramchenko@sonymobile.com>
Cc: Uladzislau Rezki <uladzislau.rezki@sony.com>
Cc: Vasily Averin <vvs@virtuozzo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Dave Chinner has mentioned that some of the xfs code would benefit from
kvmalloc support for __GFP_NOFAIL because they have allocations that
cannot fail and they do not fit into a single page.
The large part of the vmalloc implementation already complies with the
given gfp flags so there is no work for those to be done. The area and
page table allocations are an exception to that. Implement a retry loop
for those.
Add a short sleep before retrying. 1 jiffy is a completely random
timeout. Ideally the retry would wait for an explicit event - e.g. a
change to the vmalloc space change if the failure was caused by the
space fragmentation or depletion. But there are multiple different
reasons to retry and this could become much more complex. Keep the
retry simple for now and just sleep to prevent from hogging CPUs.
Link: https://lkml.kernel.org/r/20211122153233.9924-3-mhocko@kernel.org
Signed-off-by: Michal Hocko <mhocko@suse.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Ilya Dryomov <idryomov@gmail.com>
Cc: Jeff Layton <jlayton@kernel.org>
Cc: Neil Brown <neilb@suse.de>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "extend vmalloc support for constrained allocations", v2.
Based on a recent discussion with Dave and Neil [1] I have tried to
implement NOFS, NOIO, NOFAIL support for the vmalloc to make life of
kvmalloc users easier.
A requirement for NOFAIL support for kvmalloc was new to me but this
seems to be really needed by the xfs code.
NOFS/NOIO was a known and a long term problem which was hoped to be
handled by the scope API. Those scope should have been used at the
reclaim recursion boundaries both to document them and also to remove
the necessity of NOFS/NOIO constrains for all allocations within that
scope. Instead workarounds were developed to wrap a single allocation
instead (like ceph_kvmalloc).
First patch implements NOFS/NOIO support for vmalloc. The second one
adds NOFAIL support and the third one bundles all together into kvmalloc
and drops ceph_kvmalloc which can use kvmalloc directly now.
[1] http://lkml.kernel.org/r/163184741778.29351.16920832234899124642.stgit@noble.brown
This patch (of 4):
vmalloc historically hasn't supported GFP_NO{FS,IO} requests because
page table allocations do not support externally provided gfp mask and
performed GFP_KERNEL like allocations.
Since few years we have scope (memalloc_no{fs,io}_{save,restore}) APIs
to enforce NOFS and NOIO constrains implicitly to all allocators within
the scope. There was a hope that those scopes would be defined on a
higher level when the reclaim recursion boundary starts/stops (e.g.
when a lock required during the memory reclaim is required etc.). It
seems that not all NOFS/NOIO users have adopted this approach and
instead they have taken a workaround approach to wrap a single
[k]vmalloc allocation by a scope API.
These workarounds do not serve the purpose of a better reclaim recursion
documentation and reduction of explicit GFP_NO{FS,IO} usege so let's
just provide them with the semantic they are asking for without a need
for workarounds.
Add support for GFP_NOFS and GFP_NOIO to vmalloc directly. All internal
allocations already comply with the given gfp_mask. The only current
exception is vmap_pages_range which maps kernel page tables. Infer the
proper scope API based on the given gfp mask.
[sfr@canb.auug.org.au: mm/vmalloc.c needs linux/sched/mm.h]
Link: https://lkml.kernel.org/r/20211217232641.0148710c@canb.auug.org.au
Link: https://lkml.kernel.org/r/20211122153233.9924-1-mhocko@kernel.org
Link: https://lkml.kernel.org/r/20211122153233.9924-2-mhocko@kernel.org
Signed-off-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Neil Brown <neilb@suse.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ilya Dryomov <idryomov@gmail.com>
Cc: Jeff Layton <jlayton@kernel.org>
Cc: Dave Chinner <dchinner@redhat.com>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Percpu embedded first chunk allocator is the firstly option, but it
could fail on ARM64, eg,
percpu: max_distance=0x5fcfdc640000 too large for vmalloc space 0x781fefff0000
percpu: max_distance=0x600000540000 too large for vmalloc space 0x7dffb7ff0000
percpu: max_distance=0x5fff9adb0000 too large for vmalloc space 0x5dffb7ff0000
then we could get to
WARNING: CPU: 15 PID: 461 at vmalloc.c:3087 pcpu_get_vm_areas+0x488/0x838
and the system cannot boot successfully.
Let's implement page mapping percpu first chunk allocator as a fallback
to the embedding allocator to increase the robustness of the system.
Also fix a crash when both NEED_PER_CPU_PAGE_FIRST_CHUNK and
KASAN_VMALLOC enabled.
Tested on ARM64 qemu with cmdline "percpu_alloc=page".
This patch (of 3):
There are some fixed locations in the vmalloc area be reserved in
ARM(see iotable_init()) and ARM64(see map_kernel()), but for
pcpu_page_first_chunk(), it calls vm_area_register_early() and choose
VMALLOC_START as the start address of vmap area which could be
conflicted with above address, then could trigger a BUG_ON in
vm_area_add_early().
Let's choose a suit start address by traversing the vmlist.
Link: https://lkml.kernel.org/r/20210910053354.26721-1-wangkefeng.wang@huawei.com
Link: https://lkml.kernel.org/r/20210910053354.26721-2-wangkefeng.wang@huawei.com
Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Marco Elver <elver@google.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Nicholas Piggin
Song Liu
Omar Sandoval
Andrey Konovalov
Jiapeng Chong
Uladzislau Rezki (Sony)
Uladzislau Rezki
Miaohe Lin
Anshuman Khandual
Michal Hocko
Shakeel Butt
Kefeng Wang
Chen Wandun