Hi, Boris, Borislav Petkov <bp@xxxxxxxxx> writes: > On December 27, 2024 10:54:22 AM GMT+01:00, Feng Tang <feng.tang@xxxxxxxxxxxxxxxxx> wrote: >>There is a problem report that when debugging a hard-to-reproduce panic >>issue, user wanted the kernel to not reboot by adding "panic=0" in >>kernel cmdline, so that the panic context could be kept, say the panic >>was caught randomly in the mid-night, and user hoped to check it in >>the morning. GHES panic handler may overwrite that user setting and >>force to reboot after 'ghes_panic_timeout'(30) seconds. > > Why doesn't the ghes panic handler honor a panic=0 setting? It appears that I introduced the ghes_panic_timeout originally. panic() is used for software errors, while ghes is used for hardware errors. They may have different requirements. For example, it may be OK to wait forever for a software error, but it may be better to reboot the system to contain the influence of the hardware error for some hardware errors. So, we introduced another knob for that. --- Best Regards, Huang, Ying
