Functions that work on a pointer to virtual memory such as virt_to_pfn()
and users of that function such as virt_to_page() are supposed to pass a
pointer to virtual memory, ideally a (void *) or other pointer. However
since many architectures implement virt_to_pfn() as a macro, this function
becomes polymorphic and accepts both a (unsigned long) and a (void *).
If we instead implement a proper virt_to_pfn(void *addr) function the
following happens (occurred on arch/arm):
mm/nommu.c: In function 'free_page_series':
mm/nommu.c:501:50: warning: passing argument 1 of 'virt_to_pfn'
makes pointer from integer without a cast [-Wint-conversion]
struct page *page = virt_to_page(from);
Fix this with an explicit cast.
Link: https://lkml.kernel.org/r/20220630084124.691207-6-linus.walleij@linaro.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Cc: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Jason Gunthorpe <jgg@nvidia.com>
Cc: Marco Elver <elver@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Functions that work on a pointer to virtual memory such as virt_to_pfn()
and users of that function such as virt_to_page() are supposed to pass a
pointer to virtual memory, ideally a (void *) or other pointer. However
since many architectures implement virt_to_pfn() as a macro, this function
becomes polymorphic and accepts both a (unsigned long) and a (void *).
If we instead implement a proper virt_to_pfn(void *addr) function the
following happens (occurred on arch/arm):
mm/gup.c: In function '__get_user_pages_locked':
mm/gup.c:1599:49: warning: passing argument 1 of 'virt_to_pfn'
makes pointer from integer without a cast [-Wint-conversion]
pages[i] = virt_to_page(start);
Fix this with an explicit cast.
Link: https://lkml.kernel.org/r/20220630084124.691207-5-linus.walleij@linaro.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Marco Elver <elver@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Functions that work on a pointer to virtual memory such as virt_to_pfn()
and users of that function such as virt_to_page() are supposed to pass a
pointer to virtual memory, ideally a (void *) or other pointer. However
since many architectures implement virt_to_pfn() as a macro, this function
becomes polymorphic and accepts both a (unsigned long) and a (void *).
If we instead implement a proper virt_to_pfn(void *addr) function the
following happens (occurred on arch/arm):
mm/kfence/core.c:558:30: warning: passing argument 1
of 'virt_to_pfn' makes pointer from integer without a
cast [-Wint-conversion]
In one case we can refer to __kfence_pool directly (and that is a proper
(char *) pointer) and in the other call site we use an explicit cast.
Link: https://lkml.kernel.org/r/20220630084124.691207-4-linus.walleij@linaro.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Marco Elver <elver@google.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Functions that work on a pointer to virtual memory such as virt_to_pfn()
and users of that function such as virt_to_page() are supposed to pass a
pointer to virtual memory, ideally a (void *) or other pointer. However
since many architectures implement virt_to_pfn() as a macro, this function
becomes polymorphic and accepts both a (unsigned long) and a (void *).
If we instead implement a proper virt_to_pfn(void *addr) function the
following happens (occurred on arch/arm):
mm/highmem.c:153:29: warning: passing argument 1 of
'virt_to_pfn' makes pointer from integer without a
cast [-Wint-conversion]
We already have a proper void * pointer in the scope of this function
named "vaddr" so pass that instead.
Link: https://lkml.kernel.org/r/20220630084124.691207-3-linus.walleij@linaro.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Cc: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Jason Gunthorpe <jgg@nvidia.com>
Cc: Marco Elver <elver@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
In a recent change to the Arm architecture with the end goal of removing
highmem we need to convert virt_to_phys() and virt_to_pfn() to static
inline functions.
This will make them strongly typed.
However since virt_to_* is always implemented as macros they have become
polymorphic and accept both (void *) and e.g. unsigned long as arguments.
Other functions such as virt_to_page() simply wrap virt_to_pfn() and get
affected indirectly.
To be able to proceed, patch mm to use (void *) as argument to affected
functions in all instances.
This patch (of 5):
A pointer into virtual memory is represented by a (void *) not an u32, so
the compiler warns:
lib/test_free_pages.c:20:50: warning: passing argument 1
of 'virt_to_pfn' makes pointer from integer without a
cast [-Wint-conversion]
Fix this with an explicit cast.
Link: https://lkml.kernel.org/r/20220630084124.691207-1-linus.walleij@linaro.org
Link: https://lkml.kernel.org/r/20220630084124.691207-2-linus.walleij@linaro.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Cc: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Jason Gunthorpe <jgg@nvidia.com>
Cc: Marco Elver <elver@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
struct per_cpu_pages is no longer strictly local as PCP lists can be
drained remotely using a lock for protection. While the use of local_lock
works, it goes against the intent of local_lock which is for "pure CPU
local concurrency control mechanisms and not suited for inter-CPU
concurrency control" (Documentation/locking/locktypes.rst)
local_lock protects against migration between when the percpu pointer is
accessed and the pcp->lock acquired. The lock acquisition is a preemption
point so in the worst case, a task could migrate to another NUMA node and
accidentally allocate remote memory. The main requirement is to pin the
task to a CPU that is suitable for PREEMPT_RT and !PREEMPT_RT.
Replace local_lock with helpers that pin a task to a CPU, lookup the
per-cpu structure and acquire the embedded lock. It's similar to
local_lock without breaking the intent behind the API. It is not a
complete API as only the parts needed for PCP-alloc are implemented but in
theory, the generic helpers could be promoted to a general API if there
was demand for an embedded lock within a per-cpu struct with a guarantee
that the per-cpu structure locked matches the running CPU and cannot use
get_cpu_var due to RT concerns. PCP requires these semantics to avoid
accidentally allocating remote memory.
[mgorman@techsingularity.net: use pcp_spin_trylock_irqsave instead of pcpu_spin_trylock_irqsave]
Link: https://lkml.kernel.org/r/20220627084645.GA27531@techsingularity.net
Link: https://lkml.kernel.org/r/20220624125423.6126-8-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Tested-by: Yu Zhao <yuzhao@google.com>
Reviewed-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
Tested-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Tested-by: Yu Zhao <yuzhao@google.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Some setups, notably NOHZ_FULL CPUs, are too busy to handle the per-cpu
drain work queued by __drain_all_pages(). So introduce a new mechanism to
remotely drain the per-cpu lists. It is made possible by remotely locking
'struct per_cpu_pages' new per-cpu spinlocks. A benefit of this new
scheme is that drain operations are now migration safe.
There was no observed performance degradation vs. the previous scheme.
Both netperf and hackbench were run in parallel to triggering the
__drain_all_pages(NULL, true) code path around ~100 times per second. The
new scheme performs a bit better (~5%), although the important point here
is there are no performance regressions vs. the previous mechanism.
Per-cpu lists draining happens only in slow paths.
Minchan Kim tested an earlier version and reported;
My workload is not NOHZ CPUs but run apps under heavy memory
pressure so they goes to direct reclaim and be stuck on
drain_all_pages until work on workqueue run.
unit: nanosecond
max(dur) avg(dur) count(dur)
166713013 487511.77786438033 1283
From traces, system encountered the drain_all_pages 1283 times and
worst case was 166ms and avg was 487us.
The other problem was alloc_contig_range in CMA. The PCP draining
takes several hundred millisecond sometimes though there is no
memory pressure or a few of pages to be migrated out but CPU were
fully booked.
Your patch perfectly removed those wasted time.
Link: https://lkml.kernel.org/r/20220624125423.6126-7-mgorman@techsingularity.net
Signed-off-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Tested-by: Yu Zhao <yuzhao@google.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Hugh Dickins <hughd@google.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Currently the PCP lists are protected by using local_lock_irqsave to
prevent migration and IRQ reentrancy but this is inconvenient. Remote
draining of the lists is impossible and a workqueue is required and every
task allocation/free must disable then enable interrupts which is
expensive.
As preparation for dealing with both of those problems, protect the
lists with a spinlock. The IRQ-unsafe version of the lock is used
because IRQs are already disabled by local_lock_irqsave. spin_trylock
is used in combination with local_lock_irqsave() but later will be
replaced with a spin_trylock_irqsave when the local_lock is removed.
The per_cpu_pages still fits within the same number of cache lines after
this patch relative to before the series.
struct per_cpu_pages {
spinlock_t lock; /* 0 4 */
int count; /* 4 4 */
int high; /* 8 4 */
int batch; /* 12 4 */
short int free_factor; /* 16 2 */
short int expire; /* 18 2 */
/* XXX 4 bytes hole, try to pack */
struct list_head lists[13]; /* 24 208 */
/* size: 256, cachelines: 4, members: 7 */
/* sum members: 228, holes: 1, sum holes: 4 */
/* padding: 24 */
} __attribute__((__aligned__(64)));
There is overhead in the fast path due to acquiring the spinlock even
though the spinlock is per-cpu and uncontended in the common case. Page
Fault Test (PFT) running on a 1-socket reported the following results on a
1 socket machine.
5.19.0-rc3 5.19.0-rc3
vanilla mm-pcpspinirq-v5r16
Hmean faults/sec-1 869275.7381 ( 0.00%) 874597.5167 * 0.61%*
Hmean faults/sec-3 2370266.6681 ( 0.00%) 2379802.0362 * 0.40%*
Hmean faults/sec-5 2701099.7019 ( 0.00%) 2664889.7003 * -1.34%*
Hmean faults/sec-7 3517170.9157 ( 0.00%) 3491122.8242 * -0.74%*
Hmean faults/sec-8 3965729.6187 ( 0.00%) 3939727.0243 * -0.66%*
There is a small hit in the number of faults per second but given that the
results are more stable, it's borderline noise.
[akpm@linux-foundation.org: add missing local_unlock_irqrestore() on contention path]
Link: https://lkml.kernel.org/r/20220624125423.6126-6-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Tested-by: Yu Zhao <yuzhao@google.com>
Reviewed-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
Tested-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Hugh Dickins <hughd@google.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The per_cpu_pages is cache-aligned on a standard x86-64 distribution
configuration but a later patch will add a new field which would push the
structure into the next cache line. Use only one list to store THP-sized
pages on the per-cpu list. This assumes that the vast majority of
THP-sized allocations are GFP_MOVABLE but even if it was another type, it
would not contribute to serious fragmentation that potentially causes a
later THP allocation failure. Align per_cpu_pages on the cacheline
boundary to ensure there is no false cache sharing.
After this patch, the structure sizing is;
struct per_cpu_pages {
int count; /* 0 4 */
int high; /* 4 4 */
int batch; /* 8 4 */
short int free_factor; /* 12 2 */
short int expire; /* 14 2 */
struct list_head lists[13]; /* 16 208 */
/* size: 256, cachelines: 4, members: 6 */
/* padding: 32 */
} __attribute__((__aligned__(64)));
Link: https://lkml.kernel.org/r/20220624125423.6126-3-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Tested-by: Minchan Kim <minchan@kernel.org>
Acked-by: Minchan Kim <minchan@kernel.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Tested-by: Yu Zhao <yuzhao@google.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Nicolas Saenz Julienne <nsaenzju@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "Drain remote per-cpu directly", v5.
Some setups, notably NOHZ_FULL CPUs, may be running realtime or
latency-sensitive applications that cannot tolerate interference due to
per-cpu drain work queued by __drain_all_pages(). Introduce a new
mechanism to remotely drain the per-cpu lists. It is made possible by
remotely locking 'struct per_cpu_pages' new per-cpu spinlocks. This has
two advantages, the time to drain is more predictable and other unrelated
tasks are not interrupted.
This series has the same intent as Nicolas' series "mm/page_alloc: Remote
per-cpu lists drain support" -- avoid interference of a high priority task
due to a workqueue item draining per-cpu page lists. While many workloads
can tolerate a brief interruption, it may cause a real-time task running
on a NOHZ_FULL CPU to miss a deadline and at minimum, the draining is
non-deterministic.
Currently an IRQ-safe local_lock protects the page allocator per-cpu
lists. The local_lock on its own prevents migration and the IRQ disabling
protects from corruption due to an interrupt arriving while a page
allocation is in progress.
This series adjusts the locking. A spinlock is added to struct
per_cpu_pages to protect the list contents while local_lock_irq is
ultimately replaced by just the spinlock in the final patch. This allows
a remote CPU to safely. Follow-on work should allow the spin_lock_irqsave
to be converted to spin_lock to avoid IRQs being disabled/enabled in most
cases. The follow-on patch will be one kernel release later as it is
relatively high risk and it'll make bisections more clear if there are any
problems.
Patch 1 is a cosmetic patch to clarify when page->lru is storing buddy pages
and when it is storing per-cpu pages.
Patch 2 shrinks per_cpu_pages to make room for a spin lock. Strictly speaking
this is not necessary but it avoids per_cpu_pages consuming another
cache line.
Patch 3 is a preparation patch to avoid code duplication.
Patch 4 is a minor correction.
Patch 5 uses a spin_lock to protect the per_cpu_pages contents while still
relying on local_lock to prevent migration, stabilise the pcp
lookup and prevent IRQ reentrancy.
Patch 6 remote drains per-cpu pages directly instead of using a workqueue.
Patch 7 uses a normal spinlock instead of local_lock for remote draining
This patch (of 7):
The page allocator uses page->lru for storing pages on either buddy or PCP
lists. Create page->buddy_list and page->pcp_list as a union with
page->lru. This is simply to clarify what type of list a page is on in
the page allocator.
No functional change intended.
[minchan@kernel.org: fix page lru fields in macros]
Link: https://lkml.kernel.org/r/20220624125423.6126-2-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Tested-by: Minchan Kim <minchan@kernel.org>
Acked-by: Minchan Kim <minchan@kernel.org>
Reviewed-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Tested-by: Yu Zhao <yuzhao@google.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "hugetlb: speed up linear address scanning", v2.
At unmap, fork and remap time hugetlb address ranges are linearly scanned.
We can optimize these scans if the ranges are sparsely populated.
Also, enable page table "Lazy copy" for hugetlb at fork.
NOTE: Architectures not defining CONFIG_ARCH_WANT_GENERAL_HUGETLB need to
add an arch specific version hugetlb_mask_last_page() to take advantage of
sparse address scanning improvements. Baolin Wang added the routine for
arm64. Other architectures which could be optimized are: ia64, mips,
parisc, powerpc, s390, sh and sparc.
This patch (of 4):
HugeTLB address ranges are linearly scanned during fork, unmap and remap
operations. If a non-present entry is encountered, the code currently
continues to the next huge page aligned address. However, a non-present
entry implies that the page table page for that entry is not present.
Therefore, the linear scan can skip to the end of range mapped by the page
table page. This can speed operations on large sparsely populated hugetlb
mappings.
Create a new routine hugetlb_mask_last_page() that will return an address
mask. When the mask is ORed with an address, the result will be the
address of the last huge page mapped by the associated page table page.
Use this mask to update addresses in routines which linearly scan hugetlb
address ranges when a non-present pte is encountered.
hugetlb_mask_last_page is related to the implementation of huge_pte_offset
as hugetlb_mask_last_page is called when huge_pte_offset returns NULL.
This patch only provides a complete hugetlb_mask_last_page implementation
when CONFIG_ARCH_WANT_GENERAL_HUGETLB is defined. Architectures which
provide their own versions of huge_pte_offset can also provide their own
version of hugetlb_mask_last_page.
Link: https://lkml.kernel.org/r/20220621235620.291305-1-mike.kravetz@oracle.com
Link: https://lkml.kernel.org/r/20220621235620.291305-2-mike.kravetz@oracle.com
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Tested-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Acked-by: Muchun Song <songmuchun@bytedance.com>
Reported-by: kernel test robot <lkp@intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Naoya Horiguchi <naoya.horiguchi@linux.dev>
Cc: James Houghton <jthoughton@google.com>
Cc: Mina Almasry <almasrymina@google.com>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Christian Borntraeger <borntraeger@linux.ibm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Cc: Rolf Eike Beer <eike-kernel@sf-tec.de>
Cc: David Hildenbrand <david@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The page fault path checks THP eligibility with __transhuge_page_enabled()
which does the similar thing as hugepage_vma_check(), so use
hugepage_vma_check() instead.
However page fault allows DAX and !anon_vma cases, so added a new flag,
in_pf, to hugepage_vma_check() to make page fault work correctly.
The in_pf flag is also used to skip shmem and file THP for page fault
since shmem handles THP in its own shmem_fault() and file THP allocation
on fault is not supported yet.
Also remove hugepage_vma_enabled() since hugepage_vma_check() is the only
caller now, it is not necessary to have a helper function.
Link: https://lkml.kernel.org/r/20220616174840.1202070-6-shy828301@gmail.com
Signed-off-by: Yang Shi <shy828301@gmail.com>
Reviewed-by: Zach O'Keefe <zokeefe@google.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The transparent_hugepage_active() was introduced to show THP eligibility
bit in smaps in proc, smaps is the only user. But it actually does the
similar check as hugepage_vma_check() which is used by khugepaged. We
definitely don't have to maintain two similar checks, so kill
transparent_hugepage_active().
This patch also fixed the wrong behavior for VM_NO_KHUGEPAGED vmas.
Also move hugepage_vma_check() to huge_memory.c and huge_mm.h since it
is not only for khugepaged anymore.
[akpm@linux-foundation.org: check vma->vm_mm, per Zach]
[akpm@linux-foundation.org: add comment to vdso check]
Link: https://lkml.kernel.org/r/20220616174840.1202070-5-shy828301@gmail.com
Signed-off-by: Yang Shi <shy828301@gmail.com>
Reviewed-by: Zach O'Keefe <zokeefe@google.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "Cleanup transhuge_xxx helpers", v5.
This series is the follow-up of the discussion about cleaning up
transhuge_xxx helpers at
https://lore.kernel.org/linux-mm/627a71f8-e879-69a5-ceb3-fc8d29d2f7f1@suse.cz/.
THP has a bunch of helpers that do VMA sanity check for different paths,
they do the similar checks for the most callsites and have a lot duplicate
codes. And it is confusing what helpers should be used at what
conditions.
This series reorganized and cleaned up the code so that we could
consolidate all the checks into hugepage_vma_check().
The transhuge_vma_enabled(), transparent_hugepage_active() and
__transparent_hugepage_enabled() are killed by this series.
This patch (of 7):
Currently the THP flag check in hugepage_vma_check() will fallthrough if
the flag is NEVER and VM_HUGEPAGE is set. This is not a problem for now
since all the callers have the flag checked before or can't be invoked if
the flag is NEVER.
However, the following patch will call hugepage_vma_check() in more
places, for example, page fault, so this flag must be checked in
hugepge_vma_check().
Link: https://lkml.kernel.org/r/20220616174840.1202070-1-shy828301@gmail.com
Link: https://lkml.kernel.org/r/20220616174840.1202070-2-shy828301@gmail.com
Signed-off-by: Yang Shi <shy828301@gmail.com>
Reviewed-by: Zach O'Keefe <zokeefe@google.com>
Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "v14 fsdax-rmap + v11 fsdax-reflink", v2.
The patchset fsdax-rmap is aimed to support shared pages tracking for
fsdax.
It moves owner tracking from dax_assocaite_entry() to pmem device driver,
by introducing an interface ->memory_failure() for struct pagemap. This
interface is called by memory_failure() in mm, and implemented by pmem
device.
Then call holder operations to find the filesystem which the corrupted
data located in, and call filesystem handler to track files or metadata
associated with this page.
Finally we are able to try to fix the corrupted data in filesystem and do
other necessary processing, such as killing processes who are using the
files affected.
The call trace is like this:
memory_failure()
|* fsdax case
|------------
|pgmap->ops->memory_failure() => pmem_pgmap_memory_failure()
| dax_holder_notify_failure() =>
| dax_device->holder_ops->notify_failure() =>
| - xfs_dax_notify_failure()
| |* xfs_dax_notify_failure()
| |--------------------------
| | xfs_rmap_query_range()
| | xfs_dax_failure_fn()
| | * corrupted on metadata
| | try to recover data, call xfs_force_shutdown()
| | * corrupted on file data
| | try to recover data, call mf_dax_kill_procs()
|* normal case
|-------------
|mf_generic_kill_procs()
The patchset fsdax-reflink attempts to add CoW support for fsdax, and
takes XFS, which has both reflink and fsdax features, as an example.
One of the key mechanisms needed to be implemented in fsdax is CoW. Copy
the data from srcmap before we actually write data to the destination
iomap. And we just copy range in which data won't be changed.
Another mechanism is range comparison. In page cache case, readpage() is
used to load data on disk to page cache in order to be able to compare
data. In fsdax case, readpage() does not work. So, we need another
compare data with direct access support.
With the two mechanisms implemented in fsdax, we are able to make reflink
and fsdax work together in XFS.
This patch (of 14):
To easily track filesystem from a pmem device, we introduce a holder for
dax_device structure, and also its operation. This holder is used to
remember who is using this dax_device:
- When it is the backend of a filesystem, the holder will be the
instance of this filesystem.
- When this pmem device is one of the targets in a mapped device, the
holder will be this mapped device. In this case, the mapped device
has its own dax_device and it will follow the first rule. So that we
can finally track to the filesystem we needed.
The holder and holder_ops will be set when filesystem is being mounted,
or an target device is being activated.
Link: https://lkml.kernel.org/r/20220603053738.1218681-1-ruansy.fnst@fujitsu.com
Link: https://lkml.kernel.org/r/20220603053738.1218681-2-ruansy.fnst@fujitsu.com
Signed-off-by: Shiyang Ruan <ruansy.fnst@fujitsu.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dan Williams <dan.j.wiliams@intel.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Jane Chu <jane.chu@oracle.com>
Cc: Goldwyn Rodrigues <rgoldwyn@suse.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Naoya Horiguchi <naoya.horiguchi@nec.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Goldwyn Rodrigues <rgoldwyn@suse.com>
Cc: Ritesh Harjani <riteshh@linux.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Test cases such as migrate_fault and migrate_multiple, were modified to
explicit migrate from device to sys memory without the need of page
faults, when using device coherent type.
Snapshot test case updated to read memory device type first and based on
that, get the proper returned results migrate_ping_pong test case added to
test explicit migration from device to sys memory for both private and
coherent zone types.
Helpers to migrate from device to sys memory and vicerversa were also
added.
Link: https://lkml.kernel.org/r/20220715150521.18165-12-alex.sierra@amd.com
Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Alistair Popple <apopple@nvidia.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jason Gunthorpe <jgg@nvidia.com>
Cc: Jerome Glisse <jglisse@redhat.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Device Coherent type uses device memory that is coherently accesible by
the CPU. This could be shown as SP (special purpose) memory range at the
BIOS-e820 memory enumeration. If no SP memory is supported in system,
this could be faked by setting CONFIG_EFI_FAKE_MEMMAP.
Currently, test_hmm only supports two different SP ranges of at least
256MB size. This could be specified in the kernel parameter variable
efi_fake_mem. Ex. Two SP ranges of 1GB starting at 0x100000000 &
0x140000000 physical address. Ex.
efi_fake_mem=1G@0x100000000:0x40000,1G@0x140000000:0x40000
Private and coherent device mirror instances can be created in the same
probed. This is done by passing the module parameters spm_addr_dev0 &
spm_addr_dev1. In this case, it will create four instances of
device_mirror. The first two correspond to private device type, the last
two to coherent type. Then, they can be easily accessed from user space
through /dev/hmm_mirror<num_device>. Usually num_device 0 and 1 are for
private, and 2 and 3 for coherent types. If no module parameters are
passed, two instances of private type device_mirror will be created only.
Link: https://lkml.kernel.org/r/20220715150521.18165-11-alex.sierra@amd.com
Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Alistair Poppple <apopple@nvidia.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jason Gunthorpe <jgg@nvidia.com>
Cc: Jerome Glisse <jglisse@redhat.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
