smb: client: Simplify a return statement in get_smb2_acl_by_path()
Subsystems (3)
PATCH 1/1 smb: client: Simplify a return statement in get_smb2_acl_by_path()
From: Markus Elfring <[email protected]>
Date: Wed, 8 Oct 2025 21:56:34 +0200
Return an error pointer without referencing another local variable in an if branch of this function implementation.
--- fs/smb/client/smb2ops.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
fs/smb/client/smb2ops.c +1 -2
@@ -3216,9 +3216,8 @@ get_smb2_acl_by_path(struct cifs_sb_info *cifs_sb, utf16_path =cifs_convert_path_to_utf16 (path, cifs_sb);if (!utf16_path) { - rc = -ENOMEM ;free_xid (xid); -return ERR_PTR (rc); +return ERR_PTR (-ENOMEM ); } oparms = (struct cifs_open_parms ) {
Signature · 1 line
2.51.0
Steve French
Appliedmerged into cifs-2.6.git for-next
Quoted 31 lines · Markus Elfring
From: Markus Elfring <[email protected]>
Date: Wed, 8 Oct 2025 21:56:34 +0200Return an error pointer without referencing another local variable in an if branch of this function implementation.
Signed-off-by Markus Elfring <[email protected]>--- fs/smb/client/smb2ops.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)fs/smb/client/smb2ops.c +1
-3@@ -3216,9 +3216,8 @@ get_smb2_acl_by_path(struct cifs_sb_info *cifs_sb, utf16_path =cifs_convert_path_to_utf16 (path, cifs_sb);if (!utf16_path) { - rc = -ENOMEM ;free_xid (xid); -return ERR_PTR (rc); +return ERR_PTR (-ENOMEM ); } oparms = (struct cifs_open_parms ) { --2.51.0
Signature · 3 lines
Thanks, Steve
A message that is not in the archive (sent off-list or not yet received): CAH2r5mtpoLscs9sodXcRMO3-dqMDBSTR+ncExdqy4dQR=4uE8A@mail.gmail.com
Quoted 1 line
This is an example of one that is probably slightly worth it,
Thanks for such a positive indication.
Quoted 1 line
it shrinks one line of code, and also doesn't have risk,
Similar source code refinements might become also interesting and helpful.
Quoted 1 line
but at least three of the others today don't shrink and sometimes grow lines of code (and don't fix anything )
Further update candidates can be found and eventually transformed also with the help of the semantic patch language (Coccinelle software).
Quoted 1 line
so are unlikely to be worth it since they slightly increase risk of adding difficulty to stable backports of future fixes
I hope that such change resistance can be reconsidered.
Regards,
Markus
A message that is not in the archive (sent off-list or not yet received): CAH2r5mtY8--9ccnm5aYfOYJ=kEBr7=y-Z_eROKDp7A6DGnxwcA@mail.gmail.com
Quoted 1 line
Three of the four you just sent today, grow the code slightly,
How much will this observation matter?
Quoted 1 line
and don't fix anything,
Do these change suggestions indicate possible improvements according to specific design aspects?
Quoted 1 line
the remaining one could be considered
I am curious how the patch reviews will be evolving further.
Regards,
Markus
Quoted 4 lines · Markus Elfring
Quoted 2 lines
Three of the four you just sent today, grow the code slightly,
How much will this observation matter?
Remember the goals here, the priorities:
1) Fixes for bugs
2) Performance improvements
3) Features that add support for missing Linux requirements (e.g. adding O_TMPFILE or "rename_exchange support)
and then lower priority but still somewhat useful:
4) cleanup patches that remove unneeded code (although has to be balanced against how it can hurt backports of important fixes), ie cleanup that makes code *smaller* and therefore slightly easier to read and maintain
5) cleanup code that adds or updates comments clarifying confusing code
Obviously "cleanup" patches that don't shrink code are often useless (as they potentially break backports of patches for no value).
Signature · 3 lines
Thanks, Steve
As pointed out by the kernel test robot a few minutes ago, this patch would introduce a regression (uninitialized rc variable in free_xid macro), so will remove this patch from for-next.
Quoted 31 lines · Markus Elfring
From: Markus Elfring <[email protected]>
Date: Wed, 8 Oct 2025 21:56:34 +0200Return an error pointer without referencing another local variable in an if branch of this function implementation.
Signed-off-by Markus Elfring <[email protected]>--- fs/smb/client/smb2ops.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)fs/smb/client/smb2ops.c +1
-3@@ -3216,9 +3216,8 @@ get_smb2_acl_by_path(struct cifs_sb_info *cifs_sb, utf16_path =cifs_convert_path_to_utf16 (path, cifs_sb);if (!utf16_path) { - rc = -ENOMEM ;free_xid (xid); -return ERR_PTR (rc); +return ERR_PTR (-ENOMEM ); } oparms = (struct cifs_open_parms ) { --2.51.0
Signature · 3 lines
Thanks, Steve
Quoted 3 lines
As pointed out by the kernel test robot a few minutes ago, this patch would introduce a regression (uninitialized rc variable in
free_xidmacro), so will remove this patch from for-next.
Will this (clang) compiler report be reconsidered once more under other circumstances?
…>> +++ b/fs/smb/client/smb2ops.c
Quoted 11 lines
fs/smb/client/smb2ops.c get_smb2_acl_by_path(struct cifs_sb_info *cifs_sb, · line 3223
Quoted 11 lines
fs/smb/client/smb2ops.c +1
-2@@ -3216,9 +3216,8 @@ get_smb2_acl_by_path(struct cifs_sb_info *cifs_sb, utf16_path =cifs_convert_path_to_utf16 (path, cifs_sb);if (!utf16_path) { - rc = -ENOMEM ;free_xid (xid); -return ERR_PTR (rc); +return ERR_PTR (-ENOMEM ); } oparms = (struct cifs_open_parms ) {
…
Can it be that the uninitialized rc variable would not really matter for the adjusted statements in such an if branch?
Regards,
Markus
Quoted 3 lines
As pointed out by the kernel test robot a few minutes ago, this patch would introduce a regression (uninitialized rc variable in
free_xidmacro), so will remove this patch from for-next.
Is there a need to express the tracing of return values in any other ways?
https://elixir.bootlin.com/linux/v6.17.1/source/fs/smb/client/cifsproto.h#L49-L58
Regards,
Markus