Btrfs: fix reservations in btrfs_page_mkwrite

Josef fixed btrfs_page_mkwrite to properly release reserved
extents if there was an error.  But if we fail to get a reservation
and we fail to dirty the inode (for ENOSPC reasons), we'll end up
trying to release a reservation we never had.

This makes sure we only release if we were able to reserve.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
This commit is contained in:
Chris Mason 2012-01-25 13:47:40 -05:00
parent 9b23062840
commit 9998eb7034
1 changed files with 8 additions and 2 deletions

View File

@ -6401,18 +6401,23 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
unsigned long zero_start;
loff_t size;
int ret;
int reserved = 0;
u64 page_start;
u64 page_end;
ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
if (!ret)
if (!ret) {
ret = btrfs_update_time(vma->vm_file);
reserved = 1;
}
if (ret) {
if (ret == -ENOMEM)
ret = VM_FAULT_OOM;
else /* -ENOSPC, -EIO, etc */
ret = VM_FAULT_SIGBUS;
goto out;
if (reserved)
goto out;
goto out_noreserve;
}
ret = VM_FAULT_NOPAGE; /* make the VM retry the fault */
@ -6495,6 +6500,7 @@ out_unlock:
unlock_page(page);
out:
btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
out_noreserve:
return ret;
}