diff options
| author | Al Viro <[email protected]> | 2016-06-20 13:14:36 -0400 |
|---|---|---|
| committer | Anna Schumaker <[email protected]> | 2016-06-27 08:59:08 -0400 |
| commit | d20cb71dbf3487f24549ede1a8e2d67579b4632e (patch) | |
| tree | 2f602eb69205f724fe404f6f021c21044b560804 /arch/mips/kvm/mips.c | |
| parent | 1b982ea2ca398bdaeab6cf2aba459a9ca808f1f3 (diff) | |
make nfs_atomic_open() call d_drop() on all ->open_context() errors.
In "NFSv4: Move dentry instantiation into the NFSv4-specific atomic open code"
unconditional d_drop() after the ->open_context() had been removed. It had
been correct for success cases (there ->open_context() itself had been doing
dcache manipulations), but not for error ones. Only one of those (ENOENT)
got a compensatory d_drop() added in that commit, but in fact it should've
been done for all errors. As it is, the case of O_CREAT non-exclusive open
on a hashed negative dentry racing with e.g. symlink creation from another
client ended up with ->open_context() getting an error and proceeding to
call nfs_lookup(). On a hashed dentry, which would've instantly triggered
BUG_ON() in d_materialise_unique() (or, these days, its equivalent in
d_splice_alias()).
Cc: [email protected] # v3.10+
Tested-by: Oleg Drokin <[email protected]>
Signed-off-by: Al Viro <[email protected]>
Signed-off-by: Trond Myklebust <[email protected]>
Signed-off-by: Anna Schumaker <[email protected]>
Diffstat (limited to 'arch/mips/kvm/mips.c')
0 files changed, 0 insertions, 0 deletions