| Message ID | 1443819752-17091-2-git-send-email-ross.zwisler@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Fri, Oct 2, 2015 at 2:02 PM, Ross Zwisler <ross.zwisler@linux.intel.com> wrote: > This reverts commit 8346c416d17bf5b4ea1508662959bb62e73fd6a5. > > This commit did fix the issue it intended to fix, but it turns out that > the locking changes introduced by these two commits: > > commit 843172978bb9 ("dax: fix race between simultaneous faults") > commit 46c043ede471 ("mm: take i_mmap_lock in unmap_mapping_range() for DAX") > > had other issues as well, so they need to just be reverted. Wait, why introduce two points in the kernel history where we have a known uninitialized variable? I'd say fix up the revert of "mm: take i_mmap_lock in unmap_mapping_range() for DAX" to address the conflict with the fix, one less patch and keeps the stability rolling forward. -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Oct 02, 2015 at 02:11:03PM -0700, Dan Williams wrote: > On Fri, Oct 2, 2015 at 2:02 PM, Ross Zwisler > <ross.zwisler@linux.intel.com> wrote: > > This reverts commit 8346c416d17bf5b4ea1508662959bb62e73fd6a5. > > > > This commit did fix the issue it intended to fix, but it turns out that > > the locking changes introduced by these two commits: > > > > commit 843172978bb9 ("dax: fix race between simultaneous faults") > > commit 46c043ede471 ("mm: take i_mmap_lock in unmap_mapping_range() for DAX") > > > > had other issues as well, so they need to just be reverted. > > Wait, why introduce two points in the kernel history where we have a > known uninitialized variable? I'd say fix up the revert of "mm: take > i_mmap_lock in unmap_mapping_range() for DAX" to address the conflict > with the fix, one less patch and keeps the stability rolling forward. Essentially because I wasn't sure about the rules regarding reverts, if there are any. I assumed (perhaps incorrectly) that you'd want a 1:1 relationship between original commits and reverts. If it's better to not have intermediate breakage, sure, let's squash them. -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri 02-10-15 17:28:42, Ross Zwisler wrote: > On Fri, Oct 02, 2015 at 02:11:03PM -0700, Dan Williams wrote: > > On Fri, Oct 2, 2015 at 2:02 PM, Ross Zwisler > > <ross.zwisler@linux.intel.com> wrote: > > > This reverts commit 8346c416d17bf5b4ea1508662959bb62e73fd6a5. > > > > > > This commit did fix the issue it intended to fix, but it turns out that > > > the locking changes introduced by these two commits: > > > > > > commit 843172978bb9 ("dax: fix race between simultaneous faults") > > > commit 46c043ede471 ("mm: take i_mmap_lock in unmap_mapping_range() for DAX") > > > > > > had other issues as well, so they need to just be reverted. > > > > Wait, why introduce two points in the kernel history where we have a > > known uninitialized variable? I'd say fix up the revert of "mm: take > > i_mmap_lock in unmap_mapping_range() for DAX" to address the conflict > > with the fix, one less patch and keeps the stability rolling forward. > > Essentially because I wasn't sure about the rules regarding reverts, if there > are any. I assumed (perhaps incorrectly) that you'd want a 1:1 relationship > between original commits and reverts. If it's better to not have intermediate > breakage, sure, let's squash them. Well, reverts aren't any special commits after all. So if it is simple enough to just revert part of the patch that is broken, then just reverting that part is fine. Honza
diff --git a/fs/dax.c b/fs/dax.c index bcfb14b..7ae6df7 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -569,20 +569,8 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, if (!buffer_size_valid(&bh) || bh.b_size < PMD_SIZE) goto fallback; - sector = bh.b_blocknr << (blkbits - 9); - if (buffer_unwritten(&bh) || buffer_new(&bh)) { int i; - - length = bdev_direct_access(bh.b_bdev, sector, &kaddr, &pfn, - bh.b_size); - if (length < 0) { - result = VM_FAULT_SIGBUS; - goto out; - } - if ((length < PMD_SIZE) || (pfn & PG_PMD_COLOUR)) - goto fallback; - for (i = 0; i < PTRS_PER_PMD; i++) clear_pmem(kaddr + i * PAGE_SIZE, PAGE_SIZE); wmb_pmem(); @@ -635,6 +623,7 @@ int __dax_pmd_fault(struct vm_area_struct *vma, unsigned long address, result = VM_FAULT_NOPAGE; spin_unlock(ptl); } else { + sector = bh.b_blocknr << (blkbits - 9); length = bdev_direct_access(bh.b_bdev, sector, &kaddr, &pfn, bh.b_size); if (length < 0) {
This reverts commit 8346c416d17bf5b4ea1508662959bb62e73fd6a5. This commit did fix the issue it intended to fix, but it turns out that the locking changes introduced by these two commits: commit 843172978bb9 ("dax: fix race between simultaneous faults") commit 46c043ede471 ("mm: take i_mmap_lock in unmap_mapping_range() for DAX") had other issues as well, so they need to just be reverted. The list of issues in DAX after these commits (some newly introduced by the commits, some preexisting) can be found here: https://lkml.org/lkml/2015/9/25/602 Signed-off-by: Ross Zwisler <ross.zwisler@linux.intel.com> --- fs/dax.c | 13 +------------ 1 file changed, 1 insertion(+), 12 deletions(-)