* Regression: smb: chmod ignored (5.14.0-427.40.1.el9_4 vs. 5.14.0-503.15.1.el9_5)
@ 2025-01-24 23:05 Reiterer, Horst
2025-01-24 23:16 ` Steve French
2025-01-25 14:25 ` Paulo Alcantara
0 siblings, 2 replies; 4+ messages in thread
From: Reiterer, Horst @ 2025-01-24 23:05 UTC (permalink / raw)
To: linux-cifs@vger.kernel.org
Hi,
after updating from AlmaLinux 9.4 to 9.5 (https://repo.almalinux.org/vault/9.4/BaseOS/Source/Packages/kernel-5.14.0-427.40.1.el9_4.src.rpm vs. https://repo.almalinux.org/vault/9.5/BaseOS/Source/Packages/kernel-5.14.0-503.15.1.el9_5.src.rpm), chmod gets ignored by the CIFS filesystem when executed against a Windows file server unless chmod happens in another process.
Mount options as reported by mount:
rw,relatime,context=system_u:object_r:tmp_t:s0,vers=3.1.1,cache=strict,username=<username>
uid=0,noforceuid,gid=0,noforcegid,addr=<addr>,file_mode=0755,dir_mode=0755,soft
nounix,serverino,mapposix,reparse=nfs,rsize=4194304,wsize=4194304,bsize=1048576
retrans=1,echo_interval=60,actimeo=1,closetimeo=1
Reproduction (see source below):
./test <path/to/mountpoint>/<non-existent-filename> all
Unexpected mode 100555
strace:
openat(AT_FDCWD, "/mnt/test", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 3
close(3) = 0
chmod("/mnt/test", 0400) = 0
openat(AT_FDCWD, "/mnt/test", O_RDONLY) = 3
close(3) = 0
newfstatat(AT_FDCWD, "/mnt/test", {st_dev=makedev(0, 0x2b), st_ino=1407374884039565, st_mode=S_IFREG|0555, st_nlink=1, st_uid=0, st_gid=0, st_blksize=1048576, st_blocks=0, st_size=0, st_atime=17
37759300 /* 2025-01-24T23:55:00.242552200+0100 */, st_atime_nsec=242552200, st_mtime=1737759300 /* 2025-01-24T23:55:00.242552200+0100 */, st_mtime_nsec=242552200, st_ctime=1737759300 /* 2025-01-
24T23:55:00.243552200+0100 */, st_ctime_nsec=243552200}, 0) = 0
chmod("/mnt/test", 0600) = 0
newfstatat(AT_FDCWD, "/mnt/test", {st_dev=makedev(0, 0x2b), st_ino=1407374884039565, st_mode=S_IFREG|0555, st_nlink=1, st_uid=0, st_gid=0, st_blksize=1048576, st_blocks=0, st_size=0, st_atime=17
37759300 /* 2025-01-24T23:55:00.242552200+0100 */, st_atime_nsec=242552200, st_mtime=1737759300 /* 2025-01-24T23:55:00.242552200+0100 */, st_mtime_nsec=242552200, st_ctime=1737759300 /* 2025-01-
24T23:55:00.245552200+0100 */, st_ctime_nsec=245552200}, 0) = 0
write(2, "Unexpected mode 100555\n", 23Unexpected mode 100555
) = 23
exit_group(1) = ?
+++ exited with 1 +++
Workaround (executing the test in two processes):
./test <path/to/mountpoint>/<non-existent-filename> first
./test <path/to/mountpoint>/<non-existent-filename> last
strace:
openat(AT_FDCWD, "/mnt/test", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 3
close(3) = 0
chmod("/mnt/test", 0400) = 0
openat(AT_FDCWD, "/mnt/test", O_RDONLY) = 3
close(3) = 0
exit_group(0) = ?
+++ exited with 0 +++
newfstatat(AT_FDCWD, "/mnt/test", {st_dev=makedev(0, 0x2b), st_ino=1407374884039566, st_mode=S_IFREG|0555, st_nlink=1, st_uid=0, st_gid=0, st_blksize=1048576, st_blocks=0, st_size=0, st_atime=17
37759424 /* 2025-01-24T23:57:04.152930200+0100 */, st_atime_nsec=152930200, st_mtime=1737759424 /* 2025-01-24T23:57:04.152930200+0100 */, st_mtime_nsec=152930200, st_ctime=1737759424 /* 2025-01-
24T23:57:04.152930200+0100 */, st_ctime_nsec=152930200}, 0) = 0
chmod("/mnt/test", 0600) = 0
newfstatat(AT_FDCWD, "/mnt/test", {st_dev=makedev(0, 0x2b), st_ino=1407374884039566, st_mode=S_IFREG|0755, st_nlink=1, st_uid=0, st_gid=0, st_blksize=1048576, st_blocks=0, st_size=0, st_atime=17
37759424 /* 2025-01-24T23:57:04.152930200+0100 */, st_atime_nsec=152930200, st_mtime=1737759424 /* 2025-01-24T23:57:04.152930200+0100 */, st_mtime_nsec=152930200, st_ctime=1737759449 /* 2025-01-
24T23:57:29.621213500+0100 */, st_ctime_nsec=621213500}, 0) = 0
openat(AT_FDCWD, "/mnt/test", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 3
close(3) = 0
exit_group(0) = ?
+++ exited with 0 +++
Please let me know if your can reproduce the issue using the sample.
Cheers,
Horst Reiterer
test.c:
#include <fcntl.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <unistd.h>
int main(int argc, char *argv[])
{
if (argc != 3) {
fprintf(stderr, "test <path> <all|first|last>\n");
return 2;
}
char *path = argv[1];
char *mode = argv[2];
int fd;
if (strcmp(mode, "all") == 0 || strcmp(mode, "first") == 0) {
if ((fd = open(path, O_WRONLY|O_CREAT|O_TRUNC, 0666)) == -1) {
perror("open");
return 1;
}
close(fd);
if (chmod(path, 0400) == -1) {
perror("chmod");
return 1;
}
if ((fd = open(path, O_RDONLY)) == -1) {
perror("open");
return 1;
}
close(fd);
}
if (strcmp(mode, "all") == 0 || strcmp(mode, "last") == 0) {
struct stat statbuf;
if (stat(path, &statbuf) == -1) {
perror("stat");
return 1;
}
mode_t modebefore = statbuf.st_mode;
if (chmod(path, 0600) == -1) {
perror("chmod");
return 1;
}
if (stat(path, &statbuf) == -1) {
perror("stat");
return 1;
}
if (statbuf.st_mode == modebefore) {
fprintf(stderr, "Unexpected mode %o\n", statbuf.st_mode);
return 1;
}
if ((fd = open(path, O_WRONLY|O_CREAT|O_TRUNC, 0666)) == -1) {
perror("open");
return 1;
}
close(fd);
}
return 0;
}
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: Regression: smb: chmod ignored (5.14.0-427.40.1.el9_4 vs. 5.14.0-503.15.1.el9_5)
2025-01-24 23:05 Reiterer, Horst
@ 2025-01-24 23:16 ` Steve French
2025-01-25 14:25 ` Paulo Alcantara
1 sibling, 0 replies; 4+ messages in thread
From: Steve French @ 2025-01-24 23:16 UTC (permalink / raw)
To: Reiterer, Horst; +Cc: linux-cifs@vger.kernel.org
Since this kernel is presumably built on Redhat's 5.14 kernel, has it
been verified whether this fails the same way on current RHEL9?
On Fri, Jan 24, 2025 at 5:05 PM Reiterer, Horst
<horst.reiterer@fabasoft.com> wrote:
>
> Hi,
>
> after updating from AlmaLinux 9.4 to 9.5 (https://repo.almalinux.org/vault/9.4/BaseOS/Source/Packages/kernel-5.14.0-427.40.1.el9_4.src.rpm vs. https://repo.almalinux.org/vault/9.5/BaseOS/Source/Packages/kernel-5.14.0-503.15.1.el9_5.src.rpm), chmod gets ignored by the CIFS filesystem when executed against a Windows file server unless chmod happens in another process.
>
> Mount options as reported by mount:
>
> rw,relatime,context=system_u:object_r:tmp_t:s0,vers=3.1.1,cache=strict,username=<username>
> uid=0,noforceuid,gid=0,noforcegid,addr=<addr>,file_mode=0755,dir_mode=0755,soft
> nounix,serverino,mapposix,reparse=nfs,rsize=4194304,wsize=4194304,bsize=1048576
> retrans=1,echo_interval=60,actimeo=1,closetimeo=1
>
> Reproduction (see source below):
>
> ./test <path/to/mountpoint>/<non-existent-filename> all
> Unexpected mode 100555
>
> strace:
>
> openat(AT_FDCWD, "/mnt/test", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 3
> close(3) = 0
> chmod("/mnt/test", 0400) = 0
> openat(AT_FDCWD, "/mnt/test", O_RDONLY) = 3
> close(3) = 0
> newfstatat(AT_FDCWD, "/mnt/test", {st_dev=makedev(0, 0x2b), st_ino=1407374884039565, st_mode=S_IFREG|0555, st_nlink=1, st_uid=0, st_gid=0, st_blksize=1048576, st_blocks=0, st_size=0, st_atime=17
> 37759300 /* 2025-01-24T23:55:00.242552200+0100 */, st_atime_nsec=242552200, st_mtime=1737759300 /* 2025-01-24T23:55:00.242552200+0100 */, st_mtime_nsec=242552200, st_ctime=1737759300 /* 2025-01-
> 24T23:55:00.243552200+0100 */, st_ctime_nsec=243552200}, 0) = 0
> chmod("/mnt/test", 0600) = 0
> newfstatat(AT_FDCWD, "/mnt/test", {st_dev=makedev(0, 0x2b), st_ino=1407374884039565, st_mode=S_IFREG|0555, st_nlink=1, st_uid=0, st_gid=0, st_blksize=1048576, st_blocks=0, st_size=0, st_atime=17
> 37759300 /* 2025-01-24T23:55:00.242552200+0100 */, st_atime_nsec=242552200, st_mtime=1737759300 /* 2025-01-24T23:55:00.242552200+0100 */, st_mtime_nsec=242552200, st_ctime=1737759300 /* 2025-01-
> 24T23:55:00.245552200+0100 */, st_ctime_nsec=245552200}, 0) = 0
> write(2, "Unexpected mode 100555\n", 23Unexpected mode 100555
> ) = 23
> exit_group(1) = ?
> +++ exited with 1 +++
>
> Workaround (executing the test in two processes):
>
> ./test <path/to/mountpoint>/<non-existent-filename> first
> ./test <path/to/mountpoint>/<non-existent-filename> last
>
> strace:
>
> openat(AT_FDCWD, "/mnt/test", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 3
> close(3) = 0
> chmod("/mnt/test", 0400) = 0
> openat(AT_FDCWD, "/mnt/test", O_RDONLY) = 3
> close(3) = 0
> exit_group(0) = ?
> +++ exited with 0 +++
>
> newfstatat(AT_FDCWD, "/mnt/test", {st_dev=makedev(0, 0x2b), st_ino=1407374884039566, st_mode=S_IFREG|0555, st_nlink=1, st_uid=0, st_gid=0, st_blksize=1048576, st_blocks=0, st_size=0, st_atime=17
> 37759424 /* 2025-01-24T23:57:04.152930200+0100 */, st_atime_nsec=152930200, st_mtime=1737759424 /* 2025-01-24T23:57:04.152930200+0100 */, st_mtime_nsec=152930200, st_ctime=1737759424 /* 2025-01-
> 24T23:57:04.152930200+0100 */, st_ctime_nsec=152930200}, 0) = 0
> chmod("/mnt/test", 0600) = 0
> newfstatat(AT_FDCWD, "/mnt/test", {st_dev=makedev(0, 0x2b), st_ino=1407374884039566, st_mode=S_IFREG|0755, st_nlink=1, st_uid=0, st_gid=0, st_blksize=1048576, st_blocks=0, st_size=0, st_atime=17
> 37759424 /* 2025-01-24T23:57:04.152930200+0100 */, st_atime_nsec=152930200, st_mtime=1737759424 /* 2025-01-24T23:57:04.152930200+0100 */, st_mtime_nsec=152930200, st_ctime=1737759449 /* 2025-01-
> 24T23:57:29.621213500+0100 */, st_ctime_nsec=621213500}, 0) = 0
> openat(AT_FDCWD, "/mnt/test", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 3
> close(3) = 0
> exit_group(0) = ?
> +++ exited with 0 +++
>
> Please let me know if your can reproduce the issue using the sample.
>
> Cheers,
>
> Horst Reiterer
>
> test.c:
> #include <fcntl.h>
> #include <stdio.h>
> #include <stdlib.h>
> #include <string.h>
> #include <sys/stat.h>
> #include <sys/types.h>
> #include <unistd.h>
>
> int main(int argc, char *argv[])
> {
> if (argc != 3) {
> fprintf(stderr, "test <path> <all|first|last>\n");
> return 2;
> }
> char *path = argv[1];
> char *mode = argv[2];
> int fd;
> if (strcmp(mode, "all") == 0 || strcmp(mode, "first") == 0) {
> if ((fd = open(path, O_WRONLY|O_CREAT|O_TRUNC, 0666)) == -1) {
> perror("open");
> return 1;
> }
> close(fd);
> if (chmod(path, 0400) == -1) {
> perror("chmod");
> return 1;
> }
> if ((fd = open(path, O_RDONLY)) == -1) {
> perror("open");
> return 1;
> }
> close(fd);
> }
> if (strcmp(mode, "all") == 0 || strcmp(mode, "last") == 0) {
> struct stat statbuf;
> if (stat(path, &statbuf) == -1) {
> perror("stat");
> return 1;
> }
> mode_t modebefore = statbuf.st_mode;
> if (chmod(path, 0600) == -1) {
> perror("chmod");
> return 1;
> }
> if (stat(path, &statbuf) == -1) {
> perror("stat");
> return 1;
> }
> if (statbuf.st_mode == modebefore) {
> fprintf(stderr, "Unexpected mode %o\n", statbuf.st_mode);
> return 1;
> }
> if ((fd = open(path, O_WRONLY|O_CREAT|O_TRUNC, 0666)) == -1) {
> perror("open");
> return 1;
> }
> close(fd);
> }
> return 0;
> }
>
--
Thanks,
Steve
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: Regression: smb: chmod ignored (5.14.0-427.40.1.el9_4 vs. 5.14.0-503.15.1.el9_5)
2025-01-24 23:05 Reiterer, Horst
2025-01-24 23:16 ` Steve French
@ 2025-01-25 14:25 ` Paulo Alcantara
1 sibling, 0 replies; 4+ messages in thread
From: Paulo Alcantara @ 2025-01-25 14:25 UTC (permalink / raw)
To: Reiterer, Horst, linux-cifs@vger.kernel.org
Hi,
Thanks for the report.
"Reiterer, Horst" <horst.reiterer@fabasoft.com> writes:
> after updating from AlmaLinux 9.4 to 9.5
> (https://repo.almalinux.org/vault/9.4/BaseOS/Source/Packages/kernel-5.14.0-427.40.1.el9_4.src.rpm
> vs. https://repo.almalinux.org/vault/9.5/BaseOS/Source/Packages/kernel-5.14.0-503.15.1.el9_5.src.rpm),
> chmod gets ignored by the CIFS filesystem when executed against a
> Windows file server unless chmod happens in another process.
Yeah, I was able to reproduce it with mainline kernel as well.
Could you please file a bug [1] against File System -> CIFS component?
Thanks.
[1] https://bugzilla.kernel.org/
^ permalink raw reply [flat|nested] 4+ messages in thread
* RE: Regression: smb: chmod ignored (5.14.0-427.40.1.el9_4 vs. 5.14.0-503.15.1.el9_5)
@ 2025-01-25 22:22 Reiterer, Horst
0 siblings, 0 replies; 4+ messages in thread
From: Reiterer, Horst @ 2025-01-25 22:22 UTC (permalink / raw)
To: pc@manguebit.com; +Cc: linux-cifs@vger.kernel.org
Hi Paulo,
thanks for testing with mainline. Bug created:
https://bugzilla.kernel.org/show_bug.cgi?id=219724
Cheers,
Horst Reiterer
-----Ursprüngliche Nachricht-----
Von: Paulo Alcantara <pc@manguebit.com>
Gesendet: Samstag, 25. Januar 2025 15:25
An: Reiterer, Horst <horst.reiterer@fabasoft.com>; linux-cifs@vger.kernel.org
Betreff: Re: Regression: smb: chmod ignored (5.14.0-427.40.1.el9_4 vs. 5.14.0-503.15.1.el9_5)
Hi,
Thanks for the report.
"Reiterer, Horst" <horst.reiterer@fabasoft.com> writes:
> after updating from AlmaLinux 9.4 to 9.5
> (https://repo.almalinux.org/vault/9.4/BaseOS/Source/Packages/kernel-5.
> 14.0-427.40.1.el9_4.src.rpm vs.
> https://repo.almalinux.org/vault/9.5/BaseOS/Source/Packages/kernel-5.1
> 4.0-503.15.1.el9_5.src.rpm), chmod gets ignored by the CIFS filesystem
> when executed against a Windows file server unless chmod happens in
> another process.
Yeah, I was able to reproduce it with mainline kernel as well.
Could you please file a bug [1] against File System -> CIFS component?
Thanks.
[1] https://bugzilla.kernel.org/
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2025-01-25 22:22 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-01-25 22:22 Regression: smb: chmod ignored (5.14.0-427.40.1.el9_4 vs. 5.14.0-503.15.1.el9_5) Reiterer, Horst
-- strict thread matches above, loose matches on Subject: below --
2025-01-24 23:05 Reiterer, Horst
2025-01-24 23:16 ` Steve French
2025-01-25 14:25 ` Paulo Alcantara
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox