cifs: handle -EINTR in cifs_setattr
Subsystems (2)
PATCH 1/1 cifs: handle -EINTR in cifs_setattr
RHBZ: 1848178
Some calls that set attributes, like utimensat(), are not supposed to return -EINTR and thus do not have handlers for this in glibc which causes us to leak -EINTR to the applications which are also unprepared to handle it.
For example tar will break if utimensat() return -EINTR and abort unpacking the archive. Other applications may break too.
To handle this we add checks, and retry, for -EINTR in cifs_setattr()
--- fs/cifs/inode.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)
fs/cifs/inode.c +8 -4
@@ -2879,13 +2879,17 @@ cifs_setattr(struct dentry *direntry, struct iattr *attrs) {struct cifs_sb_info *cifs_sb =CIFS_SB (direntry->d_sb);struct cifs_tcon *pTcon =cifs_sb_master_tcon (cifs_sb); +int rc; -if (pTcon->unix_ext) -return cifs_setattr_unix (direntry, attrs); - -return cifs_setattr_nounix (direntry, attrs); +do { +if (pTcon->unix_ext) + rc =cifs_setattr_unix (direntry, attrs); +else + rc =cifs_setattr_nounix (direntry, attrs); + }while (rc == -EINTR );/* BB: add cifs_setattr_legacy for really old servers */ +return rc; }#if 0
Signature · 1 line
2.13.6
tentatively merged into cifs-2.6.git for-next
Quoted 46 lines · Ronnie Sahlberg
RHBZ: 1848178
Some calls that set attributes, like
utimensat(), are not supposed to return-EINTRand thus do not have handlers for this in glibc which causes us to leak-EINTRto the applications which are also unprepared to handle it.For example tar will break if
utimensat()return-EINTRand abort unpacking the archive. Other applications may break too.To handle this we add checks, and retry, for
-EINTRincifs_setattr()Signed-off-by Ronnie Sahlberg <[email protected]>--- fs/cifs/inode.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)fs/cifs/inode.c +8
-5@@ -2879,13 +2879,17 @@ cifs_setattr(struct dentry *direntry, struct iattr *attrs) {struct cifs_sb_info *cifs_sb =CIFS_SB (direntry->d_sb);struct cifs_tcon *pTcon =cifs_sb_master_tcon (cifs_sb); +int rc; -if (pTcon->unix_ext) -return cifs_setattr_unix (direntry, attrs); - -return cifs_setattr_nounix (direntry, attrs); +do { +if (pTcon->unix_ext) + rc =cifs_setattr_unix (direntry, attrs); +else + rc =cifs_setattr_nounix (direntry, attrs); + }while (rc == -EINTR );/* BB: add cifs_setattr_legacy for really old servers */ +return rc; }#if 0 --2.13.6
Signature · 3 lines
Thanks, Steve
Question: What causes cifs_setattr_* to return -EINTR?
Does an infinite reattempt make sense here? Or should we give up after N attempts?
Also, should we cifsFYI log before we re-attempt?
Regards,
Shyam
Quoted 56 lines · Steve French
tentatively merged into cifs-2.6.git for-next
Quoted 44 lines · Ronnie Sahlberg
RHBZ: 1848178
Some calls that set attributes, like
utimensat(), are not supposed to return-EINTRand thus do not have handlers for this in glibc which causes us to leak-EINTRto the applications which are also unprepared to handle it.For example tar will break if
utimensat()return-EINTRand abort unpacking the archive. Other applications may break too.To handle this we add checks, and retry, for
-EINTRincifs_setattr()Signed-off-by Ronnie Sahlberg <[email protected]>--- fs/cifs/inode.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)fs/cifs/inode.c +8
-5@@ -2879,13 +2879,17 @@ cifs_setattr(struct dentry *direntry, struct iattr *attrs) {struct cifs_sb_info *cifs_sb =CIFS_SB (direntry->d_sb);struct cifs_tcon *pTcon =cifs_sb_master_tcon (cifs_sb); +int rc; -if (pTcon->unix_ext) -return cifs_setattr_unix (direntry, attrs); - -return cifs_setattr_nounix (direntry, attrs); +do { +if (pTcon->unix_ext) + rc =cifs_setattr_unix (direntry, attrs); +else + rc =cifs_setattr_nounix (direntry, attrs); + }while (rc == -EINTR );/* BB: add cifs_setattr_legacy for really old servers */ +return rc; }#if 0 --Quoted 2 lines
2.13.6
--
Thanks,Steve
Signature · 1 line
-Shyam
Ignore my second question. I got the answer in your description.
Quoted 71 lines · Shyam Prasad N
Question: What causes cifs_setattr_* to return
-EINTR?
Does an infinite reattempt make sense here? Or should we give up after N attempts?
Also, should we cifsFYI log before we re-attempt?Regards,
ShyamQuoted 55 lines · Steve French
tentatively merged into cifs-2.6.git for-next
Quoted 44 lines · Ronnie Sahlberg
RHBZ: 1848178
Some calls that set attributes, like
utimensat(), are not supposed to return-EINTRand thus do not have handlers for this in glibc which causes us to leak-EINTRto the applications which are also unprepared to handle it.For example tar will break if
utimensat()return-EINTRand abort unpacking the archive. Other applications may break too.To handle this we add checks, and retry, for
-EINTRincifs_setattr()Signed-off-by Ronnie Sahlberg <[email protected]>--- fs/cifs/inode.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)fs/cifs/inode.c +8
-5@@ -2879,13 +2879,17 @@ cifs_setattr(struct dentry *direntry, struct iattr *attrs) {struct cifs_sb_info *cifs_sb =CIFS_SB (direntry->d_sb);struct cifs_tcon *pTcon =cifs_sb_master_tcon (cifs_sb); +int rc; -if (pTcon->unix_ext) -return cifs_setattr_unix (direntry, attrs); - -return cifs_setattr_nounix (direntry, attrs); +do { +if (pTcon->unix_ext) + rc =cifs_setattr_unix (direntry, attrs); +else + rc =cifs_setattr_nounix (direntry, attrs); + }while (rc == -EINTR );/* BB: add cifs_setattr_legacy for really old servers */ +return rc; }#if 0 --Quoted 2 lines
2.13.6
--
Thanks,Quoted 1 line
Steve
--
-Shyam
Signature · 1 line
-Shyam
ronnie sahlberg
reviewerQuoted 4 lines · Shyam Prasad N
Question: What causes cifs_setattr_* to return
-EINTR?
Does an infinite reattempt make sense here? Or should we give up after N attempts?
Good question. Probably.
I think you are right we might need to limit the retries.
One case could be where we never receive a response from the server and userspace
wants to CTRL-C the thread. It might mean we have to do N CTRL-C to end the process but N is better than infinite CTRL-C :-)
I will resend with a limit on retries.
Thanks!
Quoted 67 lines
Also, should we cifsFYI log before we re-attempt?
Regards,
ShyamQuoted 55 lines · Steve French
tentatively merged into cifs-2.6.git for-next
Quoted 44 lines · Ronnie Sahlberg
RHBZ: 1848178
Some calls that set attributes, like
utimensat(), are not supposed to return-EINTRand thus do not have handlers for this in glibc which causes us to leak-EINTRto the applications which are also unprepared to handle it.For example tar will break if
utimensat()return-EINTRand abort unpacking the archive. Other applications may break too.To handle this we add checks, and retry, for
-EINTRincifs_setattr()Signed-off-by Ronnie Sahlberg <[email protected]>--- fs/cifs/inode.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)fs/cifs/inode.c +8
-5@@ -2879,13 +2879,17 @@ cifs_setattr(struct dentry *direntry, struct iattr *attrs) {struct cifs_sb_info *cifs_sb =CIFS_SB (direntry->d_sb);struct cifs_tcon *pTcon =cifs_sb_master_tcon (cifs_sb); +int rc; -if (pTcon->unix_ext) -return cifs_setattr_unix (direntry, attrs); - -return cifs_setattr_nounix (direntry, attrs); +do { +if (pTcon->unix_ext) + rc =cifs_setattr_unix (direntry, attrs); +else + rc =cifs_setattr_nounix (direntry, attrs); + }while (rc == -EINTR );/* BB: add cifs_setattr_legacy for really old servers */ +return rc; }#if 0 --Quoted 2 lines
2.13.6
--
Thanks,Quoted 1 line
Steve
--
-Shyam