Re: [PATCH] fscrypt: Simplify handling of errors during initcall
Eric Biggers <[email protected]> Sun, 5 Jul 2026 12:11:58 -0700
| Newsgroups | org.kernel.vger.linux-fscrypt,org.kernel.vger.linux-fsdevel,org.kernel.vger.linux-kernel |
|---|---|
| Message-ID | <20260705191158.GC41916@quark> |
On Thu, Jun 18, 2026 at 05:00:30PM -0700, Eric Biggers wrote:
> Since CONFIG_FS_ENCRYPTION is a bool, not a tristate, fs/crypto/ can
> only be builtin or absent entirely; it can't be a loadable module.
> Therefore, the error code that gets returned from the fscrypt_init()
> initcall is never used. If any part of the initcall does fail, which
> should never happen, the kernel will be left in a bad state.
>
> Following the usual convention for builtin code, just panic the kernel
> if any of part of the initcall fails. This simplifies the code.
>
> This closely mirrors commit e77000ccc531 ("fsverity: simplify handling
> of errors during initcall").
>
> Signed-off-by: Eric Biggers <[email protected]>
Applied to https://git.kernel.org/pub/scm/fs/fscrypt/linux.git/log/?h=for-next
- Eric