Skip to contentSkip to conversation

virtiofsd: Add `sigreturn` to the seccomp whitelist

PATCH Applied Marc Hartmayer 5 participants 15 messages Nov 25, 2022 – Dec 1, 2022

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 virtiofsd: Add `sigreturn` to the seccomp whitelist

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

---
 tools/virtiofsd/passthrough_seccomp.c | 1 +
 1 file changed, 1 insertion(+)
tools/virtiofsd/passthrough_seccomp.c +1 -0
index 888295c073de..0033dab4939e 100644
@@ -110,6 +110,7 @@ static const int syscall_allowlist[] = {
 #endif
     SCMP_SYS(set_robust_list),
     SCMP_SYS(setxattr),
+    SCMP_SYS(sigreturn),
     SCMP_SYS(symlinkat),
     SCMP_SYS(syncfs),
     SCMP_SYS(time), /* Rarely needed, except on static builds */
Signature · 1 line
2.34.1
Reviewed-by on v1 1/1
#1 lore raw
Quoted 26 lines · Marc Hartmayer

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

---
 tools/virtiofsd/passthrough_seccomp.c | 1 +
 1 file changed, 1 insertion(+)
tools/virtiofsd/passthrough_seccomp.c +1 -1
index 888295c073de..0033dab4939e 100644
@@ -110,6 +110,7 @@ static const int syscall_allowlist[] = {
 #endif
     SCMP_SYS(set_robust_list),
     SCMP_SYS(setxattr),
+    SCMP_SYS(sigreturn),
     SCMP_SYS(symlinkat),
     SCMP_SYS(syncfs),
     SCMP_SYS(time), /* Rarely needed, except on static builds */
--

2.34.1

Quoted 5 lines

_______________________________________________
Virtio-fs mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/virtio-fs

Should we add this also in the rust version?, I see we don't have it enabled either.

Signature · 1 line
German
#2 lore raw

Am 25.11.22 um 17:32 schrieb German Maglione:

Quoted 35 lines · Marc Hartmayer
Quoted 26 lines

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

---
  tools/virtiofsd/passthrough_seccomp.c | 1 +
  1 file changed, 1 insertion(+)
tools/virtiofsd/passthrough_seccomp.c +1 -1
index 888295c073de..0033dab4939e 100644
@@ -110,6 +110,7 @@ static const int syscall_allowlist[] = {
  #endif
      SCMP_SYS(set_robust_list),
      SCMP_SYS(setxattr),
+    SCMP_SYS(sigreturn),
      SCMP_SYS(symlinkat),
      SCMP_SYS(syncfs),
      SCMP_SYS(time), /* Rarely needed, except on static builds */
--

2.34.1

Quoted 5 lines

_______________________________________________
Virtio-fs mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/virtio-fs

Quoted 2 lines

Should we add this also in the rust version?, I see we don't have it enabled either.

this is probably a good idea.

#3 lore raw
In reply to German Maglione
Quoted 27 lines · German Maglione
Quoted 26 lines · Marc Hartmayer

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

---
 tools/virtiofsd/passthrough_seccomp.c | 1 +
 1 file changed, 1 insertion(+)
tools/virtiofsd/passthrough_seccomp.c +1 -1
index 888295c073de..0033dab4939e 100644
@@ -110,6 +110,7 @@ static const int syscall_allowlist[] = {
 #endif
     SCMP_SYS(set_robust_list),
     SCMP_SYS(setxattr),
+    SCMP_SYS(sigreturn),
     SCMP_SYS(symlinkat),
     SCMP_SYS(syncfs),
     SCMP_SYS(time), /* Rarely needed, except on static builds */
--

2.34.1

Quoted 7 lines
Quoted 5 lines

_______________________________________________
Virtio-fs mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/virtio-fs

Thanks.

Quoted 3 lines

Should we add this also in the rust version?, I see we don't have it enabled either.

Yep - thanks.

Quoted 4 lines
Signature · 2 lines
-- 
German
Signature · 8 lines
Kind regards / Beste Grüße
   Marc Hartmayer

IBM Deutschland Research & Development GmbH
Vorsitzender des Aufsichtsrats: Gregor Pillen 
Geschäftsführung: David Faller
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294
#4 lore raw
Quoted 38 lines · Marc Hartmayer
Quoted 27 lines · German Maglione
Quoted 26 lines · Marc Hartmayer

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

---
 tools/virtiofsd/passthrough_seccomp.c | 1 +
 1 file changed, 1 insertion(+)
tools/virtiofsd/passthrough_seccomp.c +1 -1
index 888295c073de..0033dab4939e 100644
@@ -110,6 +110,7 @@ static const int syscall_allowlist[] = {
 #endif
     SCMP_SYS(set_robust_list),
     SCMP_SYS(setxattr),
+    SCMP_SYS(sigreturn),
     SCMP_SYS(symlinkat),
     SCMP_SYS(syncfs),
     SCMP_SYS(time), /* Rarely needed, except on static builds */
--

2.34.1

Quoted 7 lines
Quoted 5 lines

_______________________________________________
Virtio-fs mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/virtio-fs

Quoted 7 lines

Thanks.

Quoted 3 lines

Should we add this also in the rust version?, I see we don't have it enabled either.

Yep - thanks.

Could you test this MR to see if it is ok?
https://gitlab.com/virtio-fs/virtiofsd/-/merge_requests/144

Thanks,

Signature · 1 line
German
#5 lore raw
Quoted 42 lines · German Maglione
Quoted 41 lines · Marc Hartmayer
Quoted 27 lines · German Maglione
Quoted 26 lines · Marc Hartmayer

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

---
 tools/virtiofsd/passthrough_seccomp.c | 1 +
 1 file changed, 1 insertion(+)
tools/virtiofsd/passthrough_seccomp.c +1 -1
index 888295c073de..0033dab4939e 100644
@@ -110,6 +110,7 @@ static const int syscall_allowlist[] = {
 #endif
     SCMP_SYS(set_robust_list),
     SCMP_SYS(setxattr),
+    SCMP_SYS(sigreturn),
     SCMP_SYS(symlinkat),
     SCMP_SYS(syncfs),
     SCMP_SYS(time), /* Rarely needed, except on static builds */
--

2.34.1

Quoted 7 lines
Quoted 5 lines

_______________________________________________
Virtio-fs mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/virtio-fs

Thanks.

Quoted 1 line
Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

Should we add this also in the rust version?, I see we don't have it enabled either.

Yep - thanks.

Could you test this MR to see if it is ok?
https://gitlab.com/virtio-fs/virtiofsd/-/merge_requests/144

I couldn’t reproduce the problem using the Rust version - even without your patch. With your patch applied it’s (of course) still working.

Quoted 6 lines

Thanks,

Signature · 1 line
German
Signature · 8 lines
Kind regards / Beste Grüße
   Marc Hartmayer

IBM Deutschland Research & Development GmbH
Vorsitzender des Aufsichtsrats: Gregor Pillen 
Geschäftsführung: David Faller
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294
#7 lore raw
In reply to Marc Hartmayer
Quoted 4 lines · * Marc Hartmayer ([email protected]) wrote:

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

I'm curious; doesn't that mean that some signal is being delivered and you're returning? Which one?

Dave

Quoted 20 lines
---
 tools/virtiofsd/passthrough_seccomp.c | 1 +
 1 file changed, 1 insertion(+)
tools/virtiofsd/passthrough_seccomp.c +1 -0
index 888295c073de..0033dab4939e 100644
@@ -110,6 +110,7 @@ static const int syscall_allowlist[] = {
 #endif
     SCMP_SYS(set_robust_list),
     SCMP_SYS(setxattr),
+    SCMP_SYS(sigreturn),
     SCMP_SYS(symlinkat),
     SCMP_SYS(syncfs),
     SCMP_SYS(time), /* Rarely needed, except on static builds */
Signature · 1 line
2.34.1
Signature · 1 line
Dr. David Alan Gilbert / [email protected] / Manchester, UK
#8 lore raw
Quoted 8 lines · Dr. David Alan Gilbert
Quoted 4 lines · * Marc Hartmayer ([email protected]) wrote:

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

I'm curious; doesn't that mean that some signal is being delivered and you're returning? Which one?

code=0x80000000 means that the seccomp action SECCOMP_RET_KILL_PROCESS is taken => process is killed by a SIGSYS signal (31) [1].

At least, that’s my understanding of this log message.

[1] https://man7.org/linux/man-pages/man2/seccomp.2.html

[…snip…]

Quoted 3 lines
Signature · 1 line
Dr. David Alan Gilbert / [email protected] / Manchester, UK
Signature · 8 lines
Kind regards / Beste Grüße
   Marc Hartmayer

IBM Deutschland Research & Development GmbH
Vorsitzender des Aufsichtsrats: Gregor Pillen 
Geschäftsführung: David Faller
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294
#9 lore raw
Quoted 11 lines · * Marc Hartmayer ([email protected]) wrote:
Quoted 8 lines · Dr. David Alan Gilbert
Quoted 4 lines · * Marc Hartmayer ([email protected]) wrote:

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

I'm curious; doesn't that mean that some signal is being delivered and you're returning? Which one?

Quoted 6 lines

code=0x80000000 means that the seccomp action SECCOMP_RET_KILL_PROCESS is taken => process is killed by a SIGSYS signal (31) [1].

At least, that’s my understanding of this log message.

[1] https://man7.org/linux/man-pages/man2/seccomp.2.html

But isn't that the fallout rather than the cause ? i.e. seccomp is sending a SIGSYS because the process used sigreturn, my question is why did the process call sigreturn in the first place - it must have received a signal to return from?

Dave

Quoted 15 lines

[…snip…]

Quoted 3 lines
Signature · 1 line
Dr. David Alan Gilbert / [email protected] / Manchester, UK
Signature · 8 lines
Kind regards / Beste Grüße
   Marc Hartmayer

IBM Deutschland Research & Development GmbH
Vorsitzender des Aufsichtsrats: Gregor Pillen 
Geschäftsführung: David Faller
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294
Signature · 1 line
Dr. David Alan Gilbert / [email protected] / Manchester, UK
#10 lore raw

Am 29.11.22 um 10:42 schrieb Dr. David Alan Gilbert:

Quoted 15 lines · * Marc Hartmayer ([email protected]) wrote:
Quoted 14 lines
Quoted 8 lines · Dr. David Alan Gilbert
Quoted 4 lines · * Marc Hartmayer ([email protected]) wrote:

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

I'm curious; doesn't that mean that some signal is being delivered and you're returning? Which one?

code=0x80000000 means that the seccomp action SECCOMP_RET_KILL_PROCESS is taken => process is killed by a SIGSYS signal (31) [1].

Quoted 8 lines
Quoted 3 lines

At least, that’s my understanding of this log message.

[1] https://man7.org/linux/man-pages/man2/seccomp.2.html

But isn't that the fallout rather than the cause ? i.e. seccomp is sending a SIGSYS because the process used sigreturn, my question is why did the process call sigreturn in the first place - it must have received a signal to return from?

Good question. virtiofsd seems to prepare itself for

int fuse_set_signal_handlers(struct fuse_session *se)

{
    /*
     * If we used SIG_IGN instead of the do_nothing function,
     * then we would be unable to tell if we set SIG_IGN (and
     * thus should reset to SIG_DFL in fuse_remove_signal_handlers)
     * or if it was already set to SIG_IGN (and should be left
     * untouched.
     */
    if (set_one_signal_handler(SIGHUP, exit_handler, 0) == -1 ||
        set_one_signal_handler(SIGINT, exit_handler, 0) == -1 ||
        set_one_signal_handler(SIGTERM, exit_handler, 0) == -1 ||
        set_one_signal_handler(SIGPIPE, do_nothing, 0) == -1) {
        return -1;
    }

Given that rt_sigreturn was already on the seccomp list it seems to be expected that those handlers are called.

#11 lore raw

Am 29.11.22 um 10:52 schrieb Christian Borntraeger:

Quoted 47 lines

Am 29.11.22 um 10:42 schrieb Dr. David Alan Gilbert:

Quoted 15 lines · * Marc Hartmayer ([email protected]) wrote:
Quoted 14 lines
Quoted 8 lines · Dr. David Alan Gilbert
Quoted 4 lines · * Marc Hartmayer ([email protected]) wrote:

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

I'm curious; doesn't that mean that some signal is being delivered and you're returning?  Which one?

code=0x80000000 means that the seccomp action SECCOMP_RET_KILL_PROCESS is taken => process is killed by a SIGSYS signal (31) [1].

Quoted 8 lines
Quoted 3 lines

At least, that’s my understanding of this log message.

[1] https://man7.org/linux/man-pages/man2/seccomp.2.html

But isn't that the fallout rather than the cause ? i.e. seccomp is sending a SIGSYS because the process used sigreturn, my question is why did the process call sigreturn in the first place - it must have received a signal to return from?

Good question. virtiofsd seems to prepare itself for

int fuse_set_signal_handlers(struct fuse_session *se)
{
    /*
     * If we used SIG_IGN instead of the do_nothing function,
     * then we would be unable to tell if we set SIG_IGN (and
     * thus should reset to SIG_DFL in fuse_remove_signal_handlers)
     * or if it was already set to SIG_IGN (and should be left
     * untouched.
     */
    if (set_one_signal_handler(SIGHUP, exit_handler, 0) == -1 || set_one_signal_handler(SIGINT, exit_handler, 0) == -1 || set_one_signal_handler(SIGTERM, exit_handler, 0) == -1 || set_one_signal_handler(SIGPIPE, do_nothing, 0) == -1) { return -1;
    }

Quoted 2 lines

Given that rt_sigreturn was already on the seccomp list it seems to be expected that those handlers are called.

For me, it seems to happen on shutdown:

                Stack trace of thread 1:
                #0  0x000003ffc06f348a __kernel_sigreturn (linux-vdso64.so.1 + 0x48a)
                #1  0x000003ffc06f3488 __kernel_sigreturn (linux-vdso64.so.1 + 0x488)
                #2  0x000003ff9af1be96 __GI___futex_abstimed_wait_cancelable64 (libc.so.6 + 0x9be96)
                #3  0x000003ff9af211b4 __pthread_clockjoin_ex (libc.so.6 + 0xa11b4)
                #4  0x000003ff9af2106e pthread_join@GLIBC_2.2 (libc.so.6 + 0xa106e)
                #5  0x000002aa35d2fe36 fv_queue_cleanup_thread (virtiofsd + 0x2fe36)
                #6  0x000002aa35d3152c stop_all_queues (virtiofsd + 0x3152c)
                #7  0x000002aa35d2869c main (virtiofsd + 0x2869c)
                #8  0x000003ff9aeb4872 __libc_start_call_main (libc.so.6 + 0x34872)
                #9  0x000003ff9aeb4950 __libc_start_main@@GLIBC_2.34 (libc.so.6 + 0x34950)
                #10 0x000002aa35d290a0 .annobin_libvhost_user.c_end.startup (virtiofsd + 0x290a0)
#12 lore raw
Quoted 54 lines · * Christian Borntraeger ([email protected]) wrote:

Am 29.11.22 um 10:52 schrieb Christian Borntraeger:

Quoted 47 lines

Am 29.11.22 um 10:42 schrieb Dr. David Alan Gilbert:

Quoted 15 lines · * Marc Hartmayer ([email protected]) wrote:
Quoted 14 lines
Quoted 8 lines · Dr. David Alan Gilbert
Quoted 4 lines · * Marc Hartmayer ([email protected]) wrote:

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

I'm curious; doesn't that mean that some signal is being delivered and you're returning?  Which one?

code=0x80000000 means that the seccomp action SECCOMP_RET_KILL_PROCESS is taken => process is killed by a SIGSYS signal (31) [1].

Quoted 8 lines
Quoted 3 lines

At least, that’s my understanding of this log message.

[1] https://man7.org/linux/man-pages/man2/seccomp.2.html

But isn't that the fallout rather than the cause ? i.e. seccomp is sending a SIGSYS because the process used sigreturn, my question is why did the process call sigreturn in the first place - it must have received a signal to return from?

Good question. virtiofsd seems to prepare itself for

int fuse_set_signal_handlers(struct fuse_session *se)
{
    /*
     * If we used SIG_IGN instead of the do_nothing function,
     * then we would be unable to tell if we set SIG_IGN (and
     * thus should reset to SIG_DFL in fuse_remove_signal_handlers)
     * or if it was already set to SIG_IGN (and should be left
     * untouched.
     */
    if (set_one_signal_handler(SIGHUP, exit_handler, 0) == -1 || set_one_signal_handler(SIGINT, exit_handler, 0) == -1 || set_one_signal_handler(SIGTERM, exit_handler, 0) == -1 || set_one_signal_handler(SIGPIPE, do_nothing, 0) == -1) { return -1;
    }

Quoted 2 lines

Given that rt_sigreturn was already on the seccomp list it seems to be expected that those handlers are called.

For me, it seems to happen on shutdown:

Quoted 12 lines
                Stack trace of thread 1:
                #0  0x000003ffc06f348a __kernel_sigreturn (linux-vdso64.so.1 + 0x48a)
                #1  0x000003ffc06f3488 __kernel_sigreturn (linux-vdso64.so.1 + 0x488)
                #2  0x000003ff9af1be96 __GI___futex_abstimed_wait_cancelable64 (libc.so.6 + 0x9be96)
                #3  0x000003ff9af211b4 __pthread_clockjoin_ex (libc.so.6 + 0xa11b4)
                #4  0x000003ff9af2106e pthread_join@GLIBC_2.2 (libc.so.6 + 0xa106e)
                #5  0x000002aa35d2fe36 fv_queue_cleanup_thread (virtiofsd + 0x2fe36)
                #6  0x000002aa35d3152c stop_all_queues (virtiofsd + 0x3152c)
                #7  0x000002aa35d2869c main (virtiofsd + 0x2869c)
                #8  0x000003ff9aeb4872 __libc_start_call_main (libc.so.6 + 0x34872)
                #9  0x000003ff9aeb4950 __libc_start_main@@GLIBC_2.34 (libc.so.6 + 0x34950)
                #10 0x000002aa35d290a0 .annobin_libvhost_user.c_end.startup (virtiofsd + 0x290a0)

<shrug> I guess it could be a SIGCHLD or SIGPIPE or something during shutdown; I guess especially since we have the SIGPIPE registered.

Dave

Quoted 1 line
Signature · 1 line
Dr. David Alan Gilbert / [email protected] / Manchester, UK
#13 lore raw

To find out:

# perf record -e signal:signal_deliver
...wait for QEMU shutdown...
^C
# perf script
qemu-system-x86_64 2319136 [001] 1062886.415312:
signal:signal_deliver: sig=2 errno=0 code=128 sa_handler=7fc6ccfbabc0
sa_flags=14000004

sig=2 is SIGINT. This is just an example, I didn't run virtiofsd.

#14 lore raw
In reply to Christian Borntraeger
Quoted 52 lines · Christian Borntraeger

Am 29.11.22 um 10:52 schrieb Christian Borntraeger:

Quoted 47 lines

Am 29.11.22 um 10:42 schrieb Dr. David Alan Gilbert:

Quoted 15 lines · * Marc Hartmayer ([email protected]) wrote:
Quoted 14 lines
Quoted 8 lines · Dr. David Alan Gilbert
Quoted 4 lines · * Marc Hartmayer ([email protected]) wrote:

The virtiofsd currently crashes on s390x. This is because of a sigreturn system call. See audit log below:

type=SECCOMP msg=audit(1669382477.611:459): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:virtd_t:s0-s0:c0.c1023 pid=6649 comm="virtiofsd" exe="/usr/libexec/virtiofsd" sig=31 arch=80000016 syscall=119 compat=0 ip=0x3fff15f748a code=0x80000000AUID="unset" UID="root" GID="root" ARCH=s390x SYSCALL=sigreturn

I'm curious; doesn't that mean that some signal is being delivered and you're returning?  Which one?

code=0x80000000 means that the seccomp action SECCOMP_RET_KILL_PROCESS is taken => process is killed by a SIGSYS signal (31) [1].

Quoted 8 lines
Quoted 3 lines

At least, that’s my understanding of this log message.

[1] https://man7.org/linux/man-pages/man2/seccomp.2.html

But isn't that the fallout rather than the cause ? i.e. seccomp is sending a SIGSYS because the process used sigreturn, my question is why did the process call sigreturn in the first place - it must have received a signal to return from?

Good question. virtiofsd seems to prepare itself for

int fuse_set_signal_handlers(struct fuse_session *se)
{
    /*
     * If we used SIG_IGN instead of the do_nothing function,
     * then we would be unable to tell if we set SIG_IGN (and
     * thus should reset to SIG_DFL in fuse_remove_signal_handlers)
     * or if it was already set to SIG_IGN (and should be left
     * untouched.
     */
    if (set_one_signal_handler(SIGHUP, exit_handler, 0) == -1 || set_one_signal_handler(SIGINT, exit_handler, 0) == -1 || set_one_signal_handler(SIGTERM, exit_handler, 0) == -1 || set_one_signal_handler(SIGPIPE, do_nothing, 0) == -1) { return -1;
    }

Quoted 2 lines

Given that rt_sigreturn was already on the seccomp list it seems to be expected that those handlers are called.

For me, it seems to happen on shutdown:

Quoted 14 lines
                 Stack trace of thread 1:
                 #0  0x000003ffc06f348a __kernel_sigreturn (linux-vdso64.so.1 + 0x48a)
                 #1  0x000003ffc06f3488 __kernel_sigreturn (linux-vdso64.so.1 + 0x488)
                 #2  0x000003ff9af1be96 __GI___futex_abstimed_wait_cancelable64 (libc.so.6 + 0x9be96)
                 #3  0x000003ff9af211b4 __pthread_clockjoin_ex (libc.so.6 + 0xa11b4)
                 #4  0x000003ff9af2106e pthread_join@GLIBC_2.2 (libc.so.6 + 0xa106e)
                 #5  0x000002aa35d2fe36 fv_queue_cleanup_thread (virtiofsd + 0x2fe36)
                 #6  0x000002aa35d3152c stop_all_queues (virtiofsd + 0x3152c)
                 #7  0x000002aa35d2869c main (virtiofsd + 0x2869c)
                 #8  0x000003ff9aeb4872 __libc_start_call_main (libc.so.6 + 0x34872)
                 #9  0x000003ff9aeb4950 __libc_start_main@@GLIBC_2.34 (libc.so.6 + 0x34950)
                 #10 0x000002aa35d290a0 .annobin_libvhost_user.c_end.startup (virtiofsd + 0x290a0)

That’s also what I see.

--
Kind regards / Beste Grüße
Marc Hartmayer

IBM Deutschland Research & Development GmbH Vorsitzender des Aufsichtsrats: Gregor Pillen Geschäftsführung: David Faller
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294