Bug 218574
| Summary: | Setting negative PANIC_TIMEOUT values not respected | ||
|---|---|---|---|
| Product: | Other | Reporter: | Donald Carr (sirspudd) |
| Component: | Configuration | Assignee: | other_configuration (other_configuration) |
| Status: | NEW --- | ||
| Severity: | normal | CC: | sirspudd |
| Priority: | P3 | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Kernel Version: | Subsystem: | ||
| Regression: | No | Bisected commit-id: | |
despite there being existing defconfigs which specify negative values, this channel does not actually appear to work, and the kernel behaves as if the sign is being discarded. A PANIC_TIMEOUT=-1 reboots after a second, not immediately. Hardcoding the value in situ works, passing it via defconfig not so much. I assume the negative value code path is intended for consumption |
I am working against the v6.7 tag: I have spent some time trying to get panic=-1 respected in the kernel. The code was clearly there for negative values and there are consumers: ./arch/arm/configs/aspeed_g5_defconfig:CONFIG_PANIC_TIMEOUT=-1 ./arch/arm/configs/aspeed_g4_defconfig:CONFIG_PANIC_TIMEOUT=-1 ./arch/arm/configs/coltrane_defconfig:CONFIG_PANIC_TIMEOUT=-1 ./arch/mips/configs/ath25_defconfig:CONFIG_MTD_REDBOOT_DIRECTORY_BLOCK=-2 ./arch/mips/configs/ath79_defconfig:CONFIG_MTD_REDBOOT_DIRECTORY_BLOCK=-2 I have verified that explicitly setting panic_timeout = -1 as below works as intended. diff --git a/kernel/panic.c b/kernel/panic.c index 2807639aab51..f07ac734a8f6 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -404,6 +404,9 @@ void panic(const char *fmt, ...) if (!panic_blink) panic_blink = no_blink; + /* this is filthy but it gets the job done */ + panic_timeout = -1; + if (panic_timeout > 0) { I tried tracing the path from defconfig to panic_timeout: +++ b/scripts/kconfig/confdata.c @@ -273,6 +273,8 @@ static int conf_set_sym_val(struct symbol *sym, int def, int def_flags, char *p) if (sym_string_valid(sym, p)) { sym->def[def].val = xstrdup(p); sym->flags |= def_flags; + conf_warning("symbol value '%s' for %s", + sym->def[def].val, sym->name); } else { if (def != S_DEF_AUTO) conf_warning("symbol value '%s' invalid for %s", and can tell you a value of -1 is present as of this point in the code. I dont know if this is broken by design or not, but it seems unlikely given the code which handles negative values.