Re: [PATCH 5.4 1/2] ext4: avoid ext4_error()'s caused by ENOMEM in the truncate path

[Sasha Levin <sashal@kernel.org>]

[ Sasha's backport helper bot ]

Hi,

The upstream commit SHA1 provided is correct: 73c384c0cdaa8ea9ca9ef2d0cff6a25930f1648e

WARNING: Author mismatch between patch and upstream commit:
Backport author: Shaoying Xu<shaoyi@amazon.com>
Commit author: Theodore Ts'o<tytso@mit.edu>


Status in newer kernel trees:
6.12.y | Present (exact SHA1)
6.6.y | Present (exact SHA1)
6.1.y | Present (exact SHA1)
5.15.y | Present (exact SHA1)
5.10.y | Present (exact SHA1)
5.4.y | Not found

Note: The patch differs from the upstream commit:
---
1:  73c384c0cdaa8 ! 1:  ad990516172ca ext4: avoid ext4_error()'s caused by ENOMEM in the truncate path
    @@ Metadata
      ## Commit message ##
         ext4: avoid ext4_error()'s caused by ENOMEM in the truncate path
     
    +    [ Upstream commit 73c384c0cdaa8ea9ca9ef2d0cff6a25930f1648e ]
    +
         We can't fail in the truncate path without requiring an fsck.
         Add work around for this by using a combination of retry loops
         and the __GFP_NOFAIL flag.
    @@ Commit message
         Reviewed-by: Harshad Shirwadkar <harshadshirwadkar@gmail.com>
         Link: https://lore.kernel.org/r/20200507175028.15061-1-pendleton@google.com
         Signed-off-by: Theodore Ts'o <tytso@mit.edu>
    +    Stable-dep-of: c26ab35702f8 ("ext4: fix slab-use-after-free in ext4_split_extent_at()")
    +    [v5.4: resolved contextual conflict in __read_extent_tree_block]
    +    Signed-off-by: Shaoying Xu <shaoyi@amazon.com>
     
      ## fs/ext4/ext4.h ##
     @@ fs/ext4/ext4.h: enum {
    @@ fs/ext4/extents.c: ext4_force_split_extent_at(handle_t *handle, struct inode *in
     +			flags);
      }
      
    - static int
    + /*
     @@ fs/ext4/extents.c: __read_extent_tree_block(const char *function, unsigned int line,
    - {
      	struct buffer_head		*bh;
      	int				err;
    -+	gfp_t				gfp_flags = __GFP_MOVABLE | GFP_NOFS;
    + 	ext4_fsblk_t			pblk;
    ++	gfp_t                           gfp_flags = __GFP_MOVABLE | GFP_NOFS;
     +
     +	if (flags & EXT4_EX_NOFAIL)
    -+		gfp_flags |= __GFP_NOFAIL;
    ++		 gfp_flags |= __GFP_NOFAIL;
      
    + 	pblk = ext4_idx_pblock(idx);
     -	bh = sb_getblk_gfp(inode->i_sb, pblk, __GFP_MOVABLE | GFP_NOFS);
     +	bh = sb_getblk_gfp(inode->i_sb, pblk, gfp_flags);
      	if (unlikely(!bh))
---

Results of testing on various branches:

| Branch                    | Patch Apply | Build Test |
|---------------------------|-------------|------------|
| stable/linux-5.4.y        |  Success    |  Success   |