On Wed, 2026-09-09 at 12:53 -0500, Frederick Lawler wrote:
> Hi Mimi and Roberto,
>
> On Mon, Jul 27, 2026 at 05:21:46PM -0500, Frederick Lawler wrote:
> > Commit 2f845882ecd2 ("integrity: Add errno field in audit message")
> > introduced an audit log function that can take an error code. It is
> > wrapped by integrity_audit_msg() that implicitly sets the error code
> > argument to zero. The problem is that there are uses of
> > integrity_audit_msg()
> > such as ima_collect_measurement() that hide the failure cause for
> > the message.
> >
> > This series aims to convert integrity_audit_msg() call sites to
> > integrity_audit_message(), and then expose error codes based on the
> > following criteria:
> >
> > 1. The log depends on a result from earlier callee, and/or
> > 2. The caller function itself returns with that result
> >
> > ima_release_policy() is the only caller I thought made sense to keep as
> > is. The message doesn't correlate with a return code, nor does it
> > depend on a result from an earlier callee in the function.
> >
> > Signed-off-by: Frederick Lawler <[email protected]>
> > ---
> > This is probably a bit more than what was asked from v1, but as I was
> > making changes and looking at other call sites, I didn't think it made
> > sense to keep integrity_audit_msg() around anymore.
> >
> > Changes in v2:
> > - Changes from v1 are now reflected in patch 3.
> > - Replace all instances of integrity_audit_msg().
> > - Convert to patch series.
> > - Link to v1:
> > https://lore.kernel.org/r/[email protected]
> >
>
> I think this series may have fallen through the cracks between the merge
> windows. Shall I respin + rebase?
Thanks for the reminder. Let me take a look at it again, before you post
another version. It might take a few days until I get to it.
Thanks,
Mimi
