On 2021/8/20 10:39, Dan Williams wrote:
On Sun, Aug 15, 2021 at 11:04 PM Shiyang Ruan <ruansy.f...@fujitsu.com> wrote:Punch hole on a reflinked file needs dax_iomap_cow_copy() too. Otherwise, data in not aligned area will be not correct. So, add the srcmap to dax_iomap_zero() and replace memset() as dax_iomap_cow_copy(). Signed-off-by: Shiyang Ruan <ruansy.f...@fujitsu.com> Reviewed-by: Ritesh Harjani <rite...@linux.ibm.com> Reviewed-by: Darrick J. Wong <djw...@kernel.org> --- fs/dax.c | 25 +++++++++++++++---------- fs/iomap/buffered-io.c | 4 ++-- include/linux/dax.h | 3 ++- 3 files changed, 19 insertions(+), 13 deletions(-) diff --git a/fs/dax.c b/fs/dax.c index e49ba68cc7e4..91ceb518f66a 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -1198,7 +1198,8 @@ static vm_fault_t dax_pmd_load_hole(struct xa_state *xas, struct vm_fault *vmf, } #endif /* CONFIG_FS_DAX_PMD */ -s64 dax_iomap_zero(loff_t pos, u64 length, struct iomap *iomap) +s64 dax_iomap_zero(loff_t pos, u64 length, const struct iomap *iomap, + const struct iomap *srcmap) { sector_t sector = iomap_sector(iomap, pos & PAGE_MASK); pgoff_t pgoff; @@ -1220,19 +1221,23 @@ s64 dax_iomap_zero(loff_t pos, u64 length, struct iomap *iomap) if (page_aligned) rc = dax_zero_page_range(iomap->dax_dev, pgoff, 1); - else + else { rc = dax_direct_access(iomap->dax_dev, pgoff, 1, &kaddr, NULL); - if (rc < 0) { - dax_read_unlock(id); - return rc; - } - - if (!page_aligned) { - memset(kaddr + offset, 0, size); + if (rc < 0) + goto out; + if (iomap->addr != srcmap->addr) { + rc = dax_iomap_cow_copy(pos, size, PAGE_SIZE, srcmap, + kaddr);Apologies, I'm confused, why is it ok to skip zeroing here?
That was a mistake. Will be fixed in next version. -- Thanks, Ruan.
