On Tue, Dec 31, 2024 at 10:23:58AM +0100, Borislav Petkov wrote: > On Tue, Dec 31, 2024 at 02:44:48PM +0800, Feng Tang wrote: > > + /* If user hasn't specifically set panic timeout, reboot to log the error! */ > > + if (!panic_timeout && !strstr(saved_command_line, "panic=")) > > And you want to scan saved_command_line because? > > Hint: look at how other code checks panic_timeout. Thanks for the hint! IIUC, you are mentioning the set_arch_panic_timeout(). One thing is, most ARCHs' default timeout is 0, while in our case, the user will also set 'panic=0' :), so we can't easily detect if the 0 is the user-set value or the OS default one. Originally I even thought about adding a flag of 'timeout_user_changed'. Any suggestion? > > Or we want to stick the orignal patch, which doesn't change the > > original flow? > > And pile more broken stuff ontop? OK, will skip this. Thanks, Feng
