Bug 199795
| Summary: | CONFIG_STATIC_USERMODEHELPER_PATH="" causes NULL pointer dereference in do_coredump() when kernel.core_pattern is set to a pipe. | ||
|---|---|---|---|
| Product: | Process Management | Reporter: | Sergey Kvachonok (ravenexp) |
| Component: | Other | Assignee: | process_other |
| Status: | NEW --- | ||
| Severity: | low | CC: | mcgrof, slyfox |
| Priority: | P1 | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Kernel Version: | 4.16.10 | Subsystem: | |
| Regression: | No | Bisected commit-id: | |
| Attachments: |
Oops dmesg
0001-umh-always-return-error-when-helper-was-not-called.patch |
||
Created attachment 288461 [details]
0001-umh-always-return-error-when-helper-was-not-called.patch
Sent patch as https://lkml.org/lkml/2020/4/15/116 Thanks for the patch and bug report! I provided an alternative patch. Let's review on the thread I replied to, I Cc'd other possible stakeholders. |
Created attachment 276125 [details] Oops dmesg My kernel config contains: CONFIG_STATIC_USERMODEHELPER=y CONFIG_STATIC_USERMODEHELPER_PATH="" My system is using systemd, which sysctls kernel.core_pattern = |/usr/lib/systemd/systemd-coredump %P %u %g %s %t %c %h %e at the boot time. Running a SEGFAULT-ing program triggers: BUG: unable to handle kernel NULL pointer dereference at 0000000000000020 IP: do_coredump+0xba0/0xe20 Full dmesg log attached. Running sysctl kernel.core_pattern = /tmp/core.%e.%p.%h.%t makes the oops go away. I understand that my kernel was configured incorrectly for use with systemd, but it should not cause a kernel oops in any case. A helpful error message would be much better.