| Message ID | 20231212093634.2464108-1-libaokun1@huawei.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [RFC] mm/filemap: avoid buffered read/write race to read inconsistent data | expand |
On Tue 12-12-23 17:36:34, Baokun Li wrote: > The following concurrency may cause the data read to be inconsistent with > the data on disk: > > cpu1 cpu2 > ------------------------------|------------------------------ > // Buffered write 2048 from 0 > ext4_buffered_write_iter > generic_perform_write > copy_page_from_iter_atomic > ext4_da_write_end > ext4_da_do_write_end > block_write_end > __block_commit_write > folio_mark_uptodate > // Buffered read 4096 from 0 smp_wmb() > ext4_file_read_iter set_bit(PG_uptodate, folio_flags) > generic_file_read_iter i_size_write // 2048 > filemap_read unlock_page(page) > filemap_get_pages > filemap_get_read_batch > folio_test_uptodate(folio) > ret = test_bit(PG_uptodate, folio_flags) > if (ret) > smp_rmb(); > // Ensure that the data in page 0-2048 is up-to-date. > > // New buffered write 2048 from 2048 > ext4_buffered_write_iter > generic_perform_write > copy_page_from_iter_atomic > ext4_da_write_end > ext4_da_do_write_end > block_write_end > __block_commit_write > folio_mark_uptodate > smp_wmb() > set_bit(PG_uptodate, folio_flags) > i_size_write // 4096 > unlock_page(page) > > isize = i_size_read(inode) // 4096 > // Read the latest isize 4096, but without smp_rmb(), there may be > // Load-Load disorder resulting in the data in the 2048-4096 range > // in the page is not up-to-date. > copy_page_to_iter > // copyout 4096 > > In the concurrency above, we read the updated i_size, but there is no read > barrier to ensure that the data in the page is the same as the i_size at > this point, so we may copy the unsynchronized page out. Hence adding the > missing read memory barrier to fix this. > > This is a Load-Load reordering issue, which only occurs on some weak > mem-ordering architectures (e.g. ARM64, ALPHA), but not on strong > mem-ordering architectures (e.g. X86). And theoretically the problem AFAIK x86 can also reorder loads vs loads so the problem can in theory happen on x86 as well. > doesn't only happen on ext4, filesystems that call filemap_read() but > don't hold inode lock (e.g. btrfs, f2fs, ubifs ...) will have this > problem, while filesystems with inode lock (e.g. xfs, nfs) won't have > this problem. > > Cc: stable@kernel.org > Signed-off-by: Baokun Li <libaokun1@huawei.com> > --- > mm/filemap.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/mm/filemap.c b/mm/filemap.c > index 71f00539ac00..6324e2ac3e74 100644 > --- a/mm/filemap.c > +++ b/mm/filemap.c > @@ -2607,6 +2607,9 @@ ssize_t filemap_read(struct kiocb *iocb, struct iov_iter *iter, > goto put_folios; > end_offset = min_t(loff_t, isize, iocb->ki_pos + iter->count); > > + /* Ensure that the page cache within isize is updated. */ Barries have to be in pairs to work and it is a good practice to document this. So here I'd have comment like: /* * Pairs with a barrier in * block_write_end()->mark_buffer_dirty() or other page * dirtying routines like iomap_write_end() to ensure * changes to page contents are visible before we see * increased inode size. */ Honza > + smp_rmb(); > + > /* > * Once we start copying data, we don't want to be touching any > * cachelines that might be contended: > -- > 2.31.1 >
On 2023/12/12 20:41, Jan Kara wrote: > On Tue 12-12-23 17:36:34, Baokun Li wrote: >> The following concurrency may cause the data read to be inconsistent with >> the data on disk: >> >> cpu1 cpu2 >> ------------------------------|------------------------------ >> // Buffered write 2048 from 0 >> ext4_buffered_write_iter >> generic_perform_write >> copy_page_from_iter_atomic >> ext4_da_write_end >> ext4_da_do_write_end >> block_write_end >> __block_commit_write >> folio_mark_uptodate >> // Buffered read 4096 from 0 smp_wmb() >> ext4_file_read_iter set_bit(PG_uptodate, folio_flags) >> generic_file_read_iter i_size_write // 2048 >> filemap_read unlock_page(page) >> filemap_get_pages >> filemap_get_read_batch >> folio_test_uptodate(folio) >> ret = test_bit(PG_uptodate, folio_flags) >> if (ret) >> smp_rmb(); >> // Ensure that the data in page 0-2048 is up-to-date. >> >> // New buffered write 2048 from 2048 >> ext4_buffered_write_iter >> generic_perform_write >> copy_page_from_iter_atomic >> ext4_da_write_end >> ext4_da_do_write_end >> block_write_end >> __block_commit_write >> folio_mark_uptodate >> smp_wmb() >> set_bit(PG_uptodate, folio_flags) >> i_size_write // 4096 >> unlock_page(page) >> >> isize = i_size_read(inode) // 4096 >> // Read the latest isize 4096, but without smp_rmb(), there may be >> // Load-Load disorder resulting in the data in the 2048-4096 range >> // in the page is not up-to-date. >> copy_page_to_iter >> // copyout 4096 >> >> In the concurrency above, we read the updated i_size, but there is no read >> barrier to ensure that the data in the page is the same as the i_size at >> this point, so we may copy the unsynchronized page out. Hence adding the >> missing read memory barrier to fix this. >> >> This is a Load-Load reordering issue, which only occurs on some weak >> mem-ordering architectures (e.g. ARM64, ALPHA), but not on strong >> mem-ordering architectures (e.g. X86). And theoretically the problem > AFAIK x86 can also reorder loads vs loads so the problem can in theory > happen on x86 as well. According to what I read in the perfbook at the link below, Loads Reordered After Loads does not happen on x86. pdf sheet 562 corresponds to page 550, Table 15.5: Summary of Memory Ordering https://mirrors.edge.kernel.org/pub/linux/kernel/people/paulmck/perfbook/perfbook-1c.2023.06.11a.pdf >> doesn't only happen on ext4, filesystems that call filemap_read() but >> don't hold inode lock (e.g. btrfs, f2fs, ubifs ...) will have this >> problem, while filesystems with inode lock (e.g. xfs, nfs) won't have >> this problem. >> >> Cc: stable@kernel.org >> Signed-off-by: Baokun Li <libaokun1@huawei.com> >> --- >> mm/filemap.c | 3 +++ >> 1 file changed, 3 insertions(+) >> >> diff --git a/mm/filemap.c b/mm/filemap.c >> index 71f00539ac00..6324e2ac3e74 100644 >> --- a/mm/filemap.c >> +++ b/mm/filemap.c >> @@ -2607,6 +2607,9 @@ ssize_t filemap_read(struct kiocb *iocb, struct iov_iter *iter, >> goto put_folios; >> end_offset = min_t(loff_t, isize, iocb->ki_pos + iter->count); >> >> + /* Ensure that the page cache within isize is updated. */ > Barries have to be in pairs to work and it is a good practice to document > this. So here I'd have comment like: > /* > * Pairs with a barrier in > * block_write_end()->mark_buffer_dirty() or other page > * dirtying routines like iomap_write_end() to ensure > * changes to page contents are visible before we see > * increased inode size. > */ > > Honza That's a very accurate description! Thanks a lot! I will add this comment in the next version. >> + smp_rmb(); >> + >> /* >> * Once we start copying data, we don't want to be touching any >> * cachelines that might be contended: >> -- >> 2.31.1 >> Thanks!
On Tue 12-12-23 21:16:16, Baokun Li wrote: > On 2023/12/12 20:41, Jan Kara wrote: > > On Tue 12-12-23 17:36:34, Baokun Li wrote: > > > The following concurrency may cause the data read to be inconsistent with > > > the data on disk: > > > > > > cpu1 cpu2 > > > ------------------------------|------------------------------ > > > // Buffered write 2048 from 0 > > > ext4_buffered_write_iter > > > generic_perform_write > > > copy_page_from_iter_atomic > > > ext4_da_write_end > > > ext4_da_do_write_end > > > block_write_end > > > __block_commit_write > > > folio_mark_uptodate > > > // Buffered read 4096 from 0 smp_wmb() > > > ext4_file_read_iter set_bit(PG_uptodate, folio_flags) > > > generic_file_read_iter i_size_write // 2048 > > > filemap_read unlock_page(page) > > > filemap_get_pages > > > filemap_get_read_batch > > > folio_test_uptodate(folio) > > > ret = test_bit(PG_uptodate, folio_flags) > > > if (ret) > > > smp_rmb(); > > > // Ensure that the data in page 0-2048 is up-to-date. > > > > > > // New buffered write 2048 from 2048 > > > ext4_buffered_write_iter > > > generic_perform_write > > > copy_page_from_iter_atomic > > > ext4_da_write_end > > > ext4_da_do_write_end > > > block_write_end > > > __block_commit_write > > > folio_mark_uptodate > > > smp_wmb() > > > set_bit(PG_uptodate, folio_flags) > > > i_size_write // 4096 > > > unlock_page(page) > > > > > > isize = i_size_read(inode) // 4096 > > > // Read the latest isize 4096, but without smp_rmb(), there may be > > > // Load-Load disorder resulting in the data in the 2048-4096 range > > > // in the page is not up-to-date. > > > copy_page_to_iter > > > // copyout 4096 > > > > > > In the concurrency above, we read the updated i_size, but there is no read > > > barrier to ensure that the data in the page is the same as the i_size at > > > this point, so we may copy the unsynchronized page out. Hence adding the > > > missing read memory barrier to fix this. > > > > > > This is a Load-Load reordering issue, which only occurs on some weak > > > mem-ordering architectures (e.g. ARM64, ALPHA), but not on strong > > > mem-ordering architectures (e.g. X86). And theoretically the problem > > AFAIK x86 can also reorder loads vs loads so the problem can in theory > > happen on x86 as well. > > According to what I read in the /perfbook /at the link below, > > Loads Reordered After Loads does not happen on x86. > > pdf sheet 562 corresponds to page 550, > > Table 15.5: Summary of Memory Ordering > > https://mirrors.edge.kernel.org/pub/linux/kernel/people/paulmck/perfbook/perfbook-1c.2023.06.11a.pdf Indeed. I stand corrected! Thanks for the link. Honza
diff --git a/mm/filemap.c b/mm/filemap.c index 71f00539ac00..6324e2ac3e74 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -2607,6 +2607,9 @@ ssize_t filemap_read(struct kiocb *iocb, struct iov_iter *iter, goto put_folios; end_offset = min_t(loff_t, isize, iocb->ki_pos + iter->count); + /* Ensure that the page cache within isize is updated. */ + smp_rmb(); + /* * Once we start copying data, we don't want to be touching any * cachelines that might be contended:
The following concurrency may cause the data read to be inconsistent with the data on disk: cpu1 cpu2 ------------------------------|------------------------------ // Buffered write 2048 from 0 ext4_buffered_write_iter generic_perform_write copy_page_from_iter_atomic ext4_da_write_end ext4_da_do_write_end block_write_end __block_commit_write folio_mark_uptodate // Buffered read 4096 from 0 smp_wmb() ext4_file_read_iter set_bit(PG_uptodate, folio_flags) generic_file_read_iter i_size_write // 2048 filemap_read unlock_page(page) filemap_get_pages filemap_get_read_batch folio_test_uptodate(folio) ret = test_bit(PG_uptodate, folio_flags) if (ret) smp_rmb(); // Ensure that the data in page 0-2048 is up-to-date. // New buffered write 2048 from 2048 ext4_buffered_write_iter generic_perform_write copy_page_from_iter_atomic ext4_da_write_end ext4_da_do_write_end block_write_end __block_commit_write folio_mark_uptodate smp_wmb() set_bit(PG_uptodate, folio_flags) i_size_write // 4096 unlock_page(page) isize = i_size_read(inode) // 4096 // Read the latest isize 4096, but without smp_rmb(), there may be // Load-Load disorder resulting in the data in the 2048-4096 range // in the page is not up-to-date. copy_page_to_iter // copyout 4096 In the concurrency above, we read the updated i_size, but there is no read barrier to ensure that the data in the page is the same as the i_size at this point, so we may copy the unsynchronized page out. Hence adding the missing read memory barrier to fix this. This is a Load-Load reordering issue, which only occurs on some weak mem-ordering architectures (e.g. ARM64, ALPHA), but not on strong mem-ordering architectures (e.g. X86). And theoretically the problem doesn't only happen on ext4, filesystems that call filemap_read() but don't hold inode lock (e.g. btrfs, f2fs, ubifs ...) will have this problem, while filesystems with inode lock (e.g. xfs, nfs) won't have this problem. Cc: stable@kernel.org Signed-off-by: Baokun Li <libaokun1@huawei.com> --- mm/filemap.c | 3 +++ 1 file changed, 3 insertions(+)