mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Bradley Morgan <bradley.morgan@linux.dev>
To: petr.pavlu@suse.com
Cc: atomlin@atomlin.com, da.gomez@kernel.org,
	linux-kernel@vger.kernel.org, linux-modules@vger.kernel.org,
	mcgrof@kernel.org, samitolvanen@google.com
Subject: Re: [PATCH] module: Clarify purpose and description of elf_validity_cache_copy()
Date: Sat, 21 Nov 2026 01:22:26 +0000	[thread overview]
Message-ID: <A862E988-8886-430B-B32B-901723DA8B05@linux.dev> (raw)
In-Reply-To: <20261119-module-elf-validation-v1-1-69a74a84552c@suse.com>

On 19 November 2026 16:25:27 GMT, Petr Pavlu <petr.pavlu@suse.com> wrote:
>With the advent of LLM agents, a number of recent patches and bug reports
>for the module loader focus on its validation of module ELF metadata.
>However, many of them are misguided.
>
>Loading a module is a privileged operation. The ELF metadata can normally
>be trusted, just as the module code itself is expected to be correct. The
>checks in elf_validity_cache_copy() are intended to catch accidental
>corruption and basic violations of the ELF format, rather than to provide
>exhaustive validation. The only requirement is that enough of the ELF data
>is validated for the module loader to safely reach the blacklist check
>without crashing on a corrupted module file. This allows a broken module
>to
>be blacklisted and prevented from loading.
>
>Clarify the purpose of elf_validity_cache_copy() and otherwise clean up
>its
>description. Hopefully, LLM agents will take this into account and the
>number of validation-related reports will decrease.
>

thxthx

Reviewed-by: Bradley Morgan <bradley.morgan@linux.dev>


>Signed-off-by: Petr Pavlu <petr.pavlu@suse.com>
>---
> kernel/module/main.c | 37 +++++++++++++++++++++++--------------
> 1 file changed, 23 insertions(+), 14 deletions(-)
>
>diff --git a/kernel/module/main.c b/kernel/module/main.c
>index 34e0204ea37c..0c204f07ca68 100644
>--- a/kernel/module/main.c
>+++ b/kernel/module/main.c
>@@ -2421,24 +2421,33 @@ static int elf_validity_cache_strtab(struct load_info *info)
> }
> 
> /*
>- * Check userspace passed ELF module against our expectations, and cache
>- * useful variables for further processing as we go.
>+ * Check the ELF module passed in from userspace against our expectations, and
>+ * cache useful variables for further processing as we go.
>  *
>- * This does basic validity checks against section offsets and sizes, the
>- * section name string table, and the indices used for it (sh_name).
>+ * Loading a module is a privileged operation. The ELF metadata can normally be
>+ * trusted, just as the module code itself is expected to be correct. These
>+ * checks are intended to catch accidental corruption and basic violations of
>+ * the ELF format, rather than to provide exhaustive validation. The only
>+ * requirement is that enough of the ELF data is validated for the module loader
>+ * to safely reach the blacklist check without crashing on a corrupted module
>+ * file. This allows a broken module to be blacklisted and prevented from
>+ * loading.
>  *
>- * As a last step, since we're already checking the ELF sections we cache
>- * useful variables which will be used later for our convenience:
>+ * The checks cover the ELF header, section offsets and sizes, section-name and
>+ * symbol string tables, offsets into those tables, and selected
>+ * section-specific constraints.
>  *
>- * 	o pointers to section headers
>- * 	o cache the modinfo symbol section
>- * 	o cache the string symbol section
>- * 	o cache the module section
>+ * While validating, cache the following for later use:
>  *
>- * As a last step we set info->mod to the temporary copy of the module in
>- * info->hdr. The final one will be allocated in move_module(). Any
>- * modifications we make to our copy of the module will be carried over
>- * to the final minted module.
>+ * * pointers to the section headers and section-name string table
>+ * * indices of relevant sections, including .modinfo, the symbol table, the
>+ *   symbol string table, the module section, and version sections
>+ * * a pointer to the symbol string table
>+ *
>+ * Finally, set info->mod to point to the temporary module struct within
>+ * info->hdr. The final copy of the module struct will be allocated in
>+ * move_module(). Any modifications to the temporary copy will be carried over
>+ * to the final one.
>  */
> static int elf_validity_cache_copy(struct load_info *info, int flags)
> {
>
>---
>base-commit: aebb7274476403040d582c61478df8f4a7bf9cd3
>change-id: 20261119-module-elf-validation-a5a0e28e6dcb
>
>

--- Thanks!

      reply	other threads:[~2026-11-21  1:22 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-19 16:25 Petr Pavlu
2026-11-21  1:22 ` Bradley Morgan [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=A862E988-8886-430B-B32B-901723DA8B05@linux.dev \
    --to=bradley.morgan@linux.dev \
    --cc=atomlin@atomlin.com \
    --cc=da.gomez@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-modules@vger.kernel.org \
    --cc=mcgrof@kernel.org \
    --cc=petr.pavlu@suse.com \
    --cc=samitolvanen@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®