On Mon, Sep 11, 2023 at 9:19 AM Stephen Smalley stephen.smalley.work@gmail.com wrote:
On Fri, Sep 8, 2023 at 6:54 PM kernel test robot lkp@intel.com wrote:
Hi Alfred,
kernel test robot noticed the following build errors:
[auto build test ERROR on 50a510a78287c15cee644f345ef8bac8977986a7]
url: https://github.com/intel-lab-lkp/linux/commits/Alfred-Piccioni/SELinux-Check... base: 50a510a78287c15cee644f345ef8bac8977986a7 patch link: https://lore.kernel.org/r/20230906115928.3749928-1-alpic%40google.com patch subject: [PATCH V2] SELinux: Check correct permissions for FS_IOC32_* config: i386-debian-10.3 (https://download.01.org/0day-ci/archive/20230909/202309090600.NSyo7d2q-lkp@i...) compiler: gcc-12 (Debian 12.2.0-14) 12.2.0 reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20230909/202309090600.NSyo7d2q-lkp@i...)
If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags | Reported-by: kernel test robot lkp@intel.com | Closes: https://lore.kernel.org/oe-kbuild-all/202309090600.NSyo7d2q-lkp@intel.com/
All errors (new ones prefixed by >>):
security/selinux/hooks.c: In function 'selinux_file_ioctl':
security/selinux/hooks.c:3647:9: error: duplicate case value
3647 | case FS_IOC32_GETFLAGS: | ^~~~security/selinux/hooks.c:3645:9: note: previously used here 3645 | case FS_IOC_GETFLAGS: | ^~~~ security/selinux/hooks.c:3648:9: error: duplicate case value 3648 | case FS_IOC32_GETVERSION: | ^~~~ security/selinux/hooks.c:3646:9: note: previously used here 3646 | case FS_IOC_GETVERSION: | ^~~~ security/selinux/hooks.c:3654:9: error: duplicate case value 3654 | case FS_IOC32_SETFLAGS: | ^~~~ security/selinux/hooks.c:3652:9: note: previously used here 3652 | case FS_IOC_SETFLAGS: | ^~~~ security/selinux/hooks.c:3655:9: error: duplicate case value 3655 | case FS_IOC32_SETVERSION: | ^~~~ security/selinux/hooks.c:3653:9: note: previously used here 3653 | case FS_IOC_SETVERSION: | ^~~~
Not sure of the right way to fix this while addressing the original issue that this patch was intended to fix. Looking in fs/ioctl.c, I see that the some FS_IOC32 values are remapped to the corresponding FS_IOC values by the compat ioctl syscall entrypoint. Also notice this comment there:
/* RED-PEN how should LSM module know it's handling 32bit? */ error = security_file_ioctl(f.file, cmd, arg); if (error) goto out;
What is both interesting and scary is that the "RED-PEN" comment seems to go back at least as far as git, which is a lifetime these days. This has been broken for a while.
So perhaps this is a defect in LSM that needs to be addressed?
I think so. I would suggest a new security_file_ioctl_compat() hook as I often worry that flags are too easy to misuse whereas a separate hook, especially one with "_compat" at the end, are a bit more clear. The good news is that of the three LSMs that have file_ioctl hook implementations it looks like only SELinux will need a dedicated compat hook implementation (*maybe* TOMOYO, but I think it should be okay reusing the existing tomoyo_file_ioctl() implementation).