The patch titled mm: direct IO starvation improvement has been added to the -mm tree. Its filename is mm-direct-io-starvation-improvement.patch Before you just go and hit "reply", please: a) Consider who else should be cc'ed b) Prefer to cc a suitable mailing list as well c) Ideally: find the original patch on the mailing list and do a reply-to-all to that, adding suitable additional cc's *** Remember to use Documentation/SubmitChecklist when testing your code *** See http://userweb.kernel.org/~akpm/stuff/added-to-mm.txt to find out what to do about this The current -mm tree may be found at http://userweb.kernel.org/~akpm/mmotm/ ------------------------------------------------------ Subject: mm: direct IO starvation improvement From: Nick Piggin <npiggin@xxxxxxx> Direct IO can invalidate and sync a lot of pagecache pages in the mapping. A 4K direct IO will actually try to sync and/or invalidate the pagecache of the entire file, for example (which might be many GB or TB large). Improve this by doing range syncs. Also, memory no longer has to be unmapped to catch the dirty bits for syncing, as dirty bits would remain coherent due to dirty mmap accounting. This fixes the immediate DM deadlocks when doing direct IO reads to block device with a mounted filesystem, if only by papering over the problem somewhat rather than addressing the fsync starvation cases. Signed-off-by: Nick Piggin <npiggin@xxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> --- mm/filemap.c | 16 +++++----------- 1 file changed, 5 insertions(+), 11 deletions(-) diff -puN mm/filemap.c~mm-direct-io-starvation-improvement mm/filemap.c --- a/mm/filemap.c~mm-direct-io-starvation-improvement +++ a/mm/filemap.c @@ -1369,7 +1369,8 @@ generic_file_aio_read(struct kiocb *iocb goto out; /* skip atime */ size = i_size_read(inode); if (pos < size) { - retval = filemap_write_and_wait(mapping); + retval = filemap_write_and_wait_range(mapping, pos, + pos + iov_length(iov, nr_segs) - 1); if (!retval) { retval = mapping->a_ops->direct_IO(READ, iocb, iov, pos, nr_segs); @@ -2137,18 +2138,10 @@ generic_file_direct_write(struct kiocb * if (count != ocount) *nr_segs = iov_shorten((struct iovec *)iov, *nr_segs, count); - /* - * Unmap all mmappings of the file up-front. - * - * This will cause any pte dirty bits to be propagated into the - * pageframes for the subsequent filemap_write_and_wait(). - */ write_len = iov_length(iov, *nr_segs); end = (pos + write_len - 1) >> PAGE_CACHE_SHIFT; - if (mapping_mapped(mapping)) - unmap_mapping_range(mapping, pos, write_len, 0); - written = filemap_write_and_wait(mapping); + written = filemap_write_and_wait_range(mapping, pos, pos + write_len - 1); if (written) goto out; @@ -2363,7 +2356,8 @@ generic_file_buffered_write(struct kiocb * the file data here, to try to honour O_DIRECT expectations. */ if (unlikely(file->f_flags & O_DIRECT) && written) - status = filemap_write_and_wait(mapping); + status = filemap_write_and_wait_range(mapping, + pos, pos + written - 1); return written ? written : status; } _ Patches currently in -mm which might be from npiggin@xxxxxxx are linux-next.patch mm-dont-mark_page_accessed-in-fault-path.patch mm-dont-mark_page_accessed-in-shmem_fault.patch mm-invoke-oom-killer-from-page-fault.patch mm-invoke-oom-killer-from-page-fault-fix.patch mm-invoke-oom-killer-from-page-fault-fix-fix-2.patch mm-write_cache_pages-cyclic-fix.patch mm-write_cache_pages-cyclic-fix-fix.patch mm-write_cache_pages-early-loop-termination.patch mm-write_cache_pages-writepage-error-fix.patch mm-write_cache_pages-integrity-fix.patch mm-write_cache_pages-cleanups.patch mm-write_cache_pages-optimise-page-cleaning.patch mm-write_cache_pages-terminate-quickly.patch mm-write_cache_pages-more-terminate-quickly.patch mm-do_sync_mapping_range-integrity-fix.patch mm-get-rid-of-pagevec_release_nonlru.patch mm-more-likely-reclaim-madv_sequential-mappings.patch mm-vmalloc-tweak-failure-printk.patch mm-vmalloc-improve-vmallocinfo.patch mm-vmalloc-use-mutex-for-purge.patch mm-vmalloc-make-lazy-unmapping-configurable.patch fs-truncate-blocks-outside-i_size-after-o_direct-write-error.patch fs-truncate-blocks-outside-i_size-after-o_direct-write-error-fix.patch hugetlb-unsigned-ret-cannot-be-negative.patch page_fault-retry-with-nopage_retry.patch page_fault-retry-with-nopage_retry-fix.patch page_fault-retry-with-nopage_retry-fix-fix.patch mm-direct-io-starvation-improvement.patch fs-remove-wb_sync_hold.patch fs-sync_sb_inodes-fix.patch fs-sys_sync-fix.patch radix-tree-gang-set-if-tagged-operation.patch mm-fsync-livelock-avoidance.patch reiser4.patch -- To unsubscribe from this list: send the line "unsubscribe mm-commits" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html