Re: WARNING: CPU: 1 PID: 83 at arch/x86/kernel/cpu/sgx/main.c:446 ksgxd+0x1b7/0x1d0

From: Jarkko Sakkinen
Date: Thu Aug 25 2022 - 01:25:16 EST


On Thu, Aug 25, 2022 at 07:57:30AM +0300, Jarkko Sakkinen wrote:
> On Fri, Aug 19, 2022 at 11:28:24AM -0700, Dave Hansen wrote:
> > On 8/19/22 09:02, Paul Menzel wrote:
> > > On the Dell XPS 13 9370, Linux 5.18.16 prints the warning below:
> > >
> > > ```
> > > [    0.000000] Linux version 5.18.0-4-amd64
> > > (debian-kernel@xxxxxxxxxxxxxxxx) (gcc-11 (Debian 11.3.0-5) 11.3.0, GNU
> > > ld (GNU Binutils for Debian) 2.38.90.20220713) #1 SMP PREEMPT_DYNAMIC
> > > Debian 5.18.16-1 (2022-08-10)
> > > [    0.000000] Command line: BOOT_IMAGE=/vmlinuz-5.18.0-4-amd64
> > > root=UUID=56f398e0-1e25-4fda-aa9f-611dece4b333 ro quiet
> > > […]
> > > [    0.000000] DMI: Dell Inc. XPS 13 9370/0RMYH9, BIOS 1.21.0 07/06/2022
> > > […]
> > > [    0.235418] sgx: EPC section 0x40200000-0x45f7ffff
> >
> > Hi Paul,
> >
> > Would you be able to send the entire dmesg, along with:
> >
> > cat /proc/iomem # (as root)
> > and
> > cpuid -1 --raw
> >
> > I'm suspecting either a BIOS problem. Reinette (cc'd) also thought this
> > might be a case of the SGX initialization getting a bit too far along
> > when it should have been disabled.
> >
> > We had some bugs where we didn't stop fast enough after spitting out the
> > "SGX Launch Control is locked..." errors.
>
> For some reason the pages do not get properly sanitized:
>
> /* sanity check: */
> WARN_ON(!list_empty(&sgx_dirty_page_list));
>
> EPC should be good, given that EREMOVE does not fail.
> If SGX would be disabled, also EREMOVE should fail.

Sorry forgot that in no circumstances we're printing the
error code inside __sgx_sanitize_pages(). I wrote a quick
patch to address this (attached) [*].

Paul,

Any chance to try the patch out? It's pretty hard to attach
e.g. kprobe to grab this info. Does it reproduce every single
time?

Alternatively: what kind of workload is triggering this?
I do own 2020 model XPS13, which might be able to
reproduce the same issue.

[*] Also: https://lore.kernel.org/linux-sgx/20220825051827.246698-1-jarkko@xxxxxxxxxx/T/#u

BR, Jarkko