[PATCH] ima: don't clear IMA_DIGSIG flag when setting non-IMA xattr

[Coiby Xu]

On Mon, Sep 15, 2025 at 12:06:14PM +0800, Coiby Xu wrote:
>On Wed, Sep 10, 2025 at 08:21:33AM -0400, Mimi Zohar wrote:
>>On Wed, 2025-09-10 at 09:36 +0800, Coiby Xu wrote:
>>>On Mon, Sep 08, 2025 at 04:58:05PM -0400, Mimi Zohar wrote:
>>>> On Mon, 2025-09-08 at 10:53 -0400, Mimi Zohar wrote:
>>>> > Hi Coiby,
>>>> >
>>>> > On Mon, 2025-09-08 at 19:12 +0800, Coiby Xu wrote:
>>>> > > >
>>>> > > > Even without an IMA appraise policy, the security xattrs are written out to the
>>>> > > > filesystem, but the IMA_DIGSIG flag is not cached.
>>>> > >
>>>> > > It seems I miss some context for the above sentence. If no IMA policy is
>>>> > > configured, no ima_iint_cache will be created. If you mean non-appraisal
>>>> > > policy, will not caching IMA_DIGSIG flag cause any problem?
>>>> >
>>>> > Sorry.  What I was trying to say is that your test program illustrates the
>>>> > problem both with or without any of the boot command line options as you
>>>> > suggested - "ima_appraise=fix evm=fix ima_policy=appraise_tcb".  Writing some
>>>> > other security xattr is a generic problem, whether the file is in policy or not,
>>>> > whether IMA or EVM are in fix mode or not.  The rpm-plugin-ima should install
>>>> > the IMA signature regardless.
>>>>
>>>> My mistake.  An appraise policy indeed needs to be defined for the file
>>>> signature to be replaced with a file hash.
>>>
>>>Thanks for the clarification! rpm-plugin-ima does try to install IMA
>>>signature as shown from the following strace output,
>>
>>Agreed. I was referring to the SELinux label, which would be installed for new
>>files, but not necessarily re-installed on existing files.  The test program
>>simplified testing.  Thank you.
>
>My pleasure! Note reinstalling a package using dnf/rpm is equivalent to
>installing a new package in terms of this issue. Because according to
>the strace output and rpm's source code, when reinstalling a package,
>the following steps happens, taking lnstat as an example,
>
>1. A temporary file "lnstat;68aee3f4" is created
>2. Read the content from RPM and write it to lnstat;68aee3f4
>3. Set file permission
>4. Set security.ima by rpm-plugin-ima
>5. Set security.selinux by rpm-plugin-selinux
>6. Rename "lnstat;68aee3f4" to lnstat
>
>And here's the strace output,
>
>    # strace rpm --reinstall ip*.rpm
>    openat(11, "lnstat;68aee3f4", O_WRONLY|O_CREAT|O_EXCL, 0200) = 12
>    dup(12)                                 = 13
>    write(13, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\0'\0\0\0\0\0\0"..., 19256) = 19256
>    close(13)                               = 0
>    getuid()                                = 0
>    fchown(12, 0, 0)                        = 0
>    fchmod(12, 0755)                        = 0
>    getuid()                                = 0
>    utimensat(12, NULL, [{tv_sec=1734480000, tv_nsec=0} /* 2024-12-17T19:00:00-0500 */, {tv_sec=1734480000, tv_nsec=0} /* 2024-12-17T19:00:00-0500 */], 0) = 0
>    fsetxattr(12, "security.ima", "\3\2\4\3232\4I\0f0d\0020O\231\341q\323Q\322\235\341\7\323\224\205\2104\24\241\331#"..., 111, 0) = 0
>    fsetxattr(12, "security.selinux", "system_u:object_r:bin_t:s0", 27, 0) = 0
>    close(12)                               = 0
>    ...
>    renameat(11, "lnstat;68aee3f4", 11, "lnstat") = 0

Btw, I realize my commit message that says the problem happens when reinstalling a
package can be a bit misleading. So in v3, I rephrase it as 
"... installing/reinstalling a package will not make good reference IMA generated"