On 10/22/25 1:28 AM, Petr Pavlu wrote:
> The TAINT_RANDSTRUCT and TAINT_FWCTL flags are mistakenly set in the
> taint_flags table as per-module flags. While this can be trivially
Ah, I had just noticed these this morning.
> corrected, the issue can be avoided altogether by removing the
> taint_flag.module field.
>
> This is possible because, since commit 7fd8329ba502 ("taint/module: Clean
> up global and module taint flags handling") in 2016, the handling of module
> taint flags has been fully generic. Specifically, module_flags_taint() can
> print all flags, and the required output buffer size is properly defined in
> terms of TAINT_FLAGS_COUNT. The actual per-module flags are always those
> added to module.taints by calls to add_taint_module().
>
> Signed-off-by: Petr Pavlu <[email protected]>
Reviewed-by: Randy Dunlap <[email protected]>
> ---
> The patch is based on linux-next (20251021) because I wanted to avoid
> a conflict with "taint: add reminder about updating docs and scripts" [1],
> which is currently queued in mm-nonmm-unstable.
>
> [1]
> https://lore.kernel.org/all/[email protected]/
>
> ---
> include/linux/panic.h | 1 -
> kernel/module/main.c | 2 +-
> kernel/panic.c | 46 ++++++++++++++++++++-----------------------
> 3 files changed, 22 insertions(+), 27 deletions(-)
>
--
~Randy
