PROBLEM: boot hang on ASUS P2B-F (regression)

PROBLEM: boot hang on ASUS P2B-F (regression)

[Nick Bowler]

[-- Attachment #1: Type: text/plain, Size: 1153 bytes --]

Hi,

After updating my PC (ASUS P2B-F w/ Pentium 2 CPU) to the latest versions,
Linux now just hangs early on boot.  The last messages printed are:

  clocksource: tsc-early: mask: 0xffffffffffffffff max_cycles: 0x5c77f5216e, max_idle_ns: 440795213641 ns
  Calibrating delay loop (skipped), value calculated using timer frequency.. 801.87 BogoMIPS (lpj=4009380)

after which there is no further output or any indication of anything
happening at all.  (Complete log is attached, gzipped).

This is a regression from 6.17, which boots successfully.  Bisection
implicates the following:

  7eee1e92684507f64ec6a75fecbd27e37174b888 is the first bad commit
  commit 7eee1e92684507f64ec6a75fecbd27e37174b888 (HEAD)
  Author: Yazen Ghannam <yazen.ghannam@amd.com>
  Date:   Mon Sep 8 15:40:34 2025 +0000
  
      x86/mce: Separate global and per-CPU quirks

This does not revert cleanly on top of 7.0-rc4.  However, the only
conflicts are in arch/x86/kernel/cpu/mce/amd.c, so I tried reverting
just the changes to core.c and intel.c, and this results in a kernel
that boots without any obvious problems.

Let me know if you need any more info!

Thanks,
  Nick

[-- Attachment #2: p2dmesg.gz --]
[-- Type: application/gzip, Size: 1826 bytes --]

Re: PROBLEM: boot hang on ASUS P2B-F (regression)

[Borislav Petkov]

On Sat, Mar 21, 2026 at 04:26:52PM -0400, Nick Bowler wrote:
> After updating my PC (ASUS P2B-F w/ Pentium 2 CPU) to the latest versions,

Pentium 2 huh?

Don't tell me you're really using that machine... :-)

Anyway, pls send .config, a dmesg from a successfully booting kernel and
/proc/cpuinfo.

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: PROBLEM: boot hang on ASUS P2B-F (regression)

[Nick Bowler]

[-- Attachment #1: Type: text/plain, Size: 961 bytes --]

On Sat, Mar 21, 2026 at 09:43:21PM +0100, Borislav Petkov wrote:
> Anyway, pls send .config, a dmesg from a successfully booting kernel and
> /proc/cpuinfo.

config and dmesg attached (both gzipped) from the successful boot of
7.0-rc4 after partially reverting the culprit.

  % cat /proc/cpuinfo
  processor	: 0
  vendor_id	: GenuineIntel
  cpu family	: 6
  model		: 5
  model name	: Pentium II (Deschutes)
  stepping	: 2
  microcode	: 0x2a
  cpu MHz		: 400.938
  cache size	: 512 KB
  fdiv_bug	: no
  f00f_bug	: no
  coma_bug	: no
  fpu		: yes
  fpu_exception	: yes
  cpuid level	: 2
  wp		: yes
  flags		: fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pse36 mmx fxsr cpuid
  bugs		: cpu_meltdown spectre_v1 spectre_v2 spec_store_bypass l1tf mds swapgs itlb_multihit spectre_v2_user old_microcode
  bogomips	: 801.87
  clflush size	: 32
  cache_alignment	: 32
  address sizes	: 36 bits physical, 32 bits virtual
  power management:

Thanks,
  Nick

[-- Attachment #2: config.gz --]
[-- Type: application/gzip, Size: 20681 bytes --]

[-- Attachment #3: dmesg_ok.gz --]
[-- Type: application/gzip, Size: 8266 bytes --]

Re: PROBLEM: boot hang on ASUS P2B-F (regression)

[Borislav Petkov]

On Sat, Mar 21, 2026 at 05:00:28PM -0400, Nick Bowler wrote:
> On Sat, Mar 21, 2026 at 09:43:21PM +0100, Borislav Petkov wrote:
> > Anyway, pls send .config, a dmesg from a successfully booting kernel and
> > /proc/cpuinfo.
> 
> config and dmesg attached (both gzipped) from the successful boot of
> 7.0-rc4 after partially reverting the culprit.

I can't reproduce in qemu here because I don't think anyone has your hardware
anymore. I'd throw it in the trash if I were you. :-P

Anyway, since you're partially reverting, can you pls try reverting the
Intel-relevant hunks and narrow it down, if possible.

From the looks of it, I'm thinking:

* comment out the intel_apply_cpu_quirks() call

* comment out this hunk

        /*
         * There are also broken BIOSes on some Pentium M and
         * earlier systems:
         */
        if (c->x86_vfm < INTEL_CORE_YONAH && mca_cfg.bootlog < 0)
                mca_cfg.bootlog = 0;

in intel_apply_global_quirks()

as those two should be relevant to your machine.

That might tell us where to look.

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Re: PROBLEM: boot hang on ASUS P2B-F (regression)

[Nick Bowler]

On Sat, Mar 21, 2026 at 10:52:24PM +0100, Borislav Petkov wrote:
> Anyway, since you're partially reverting, can you pls try reverting the
> Intel-relevant hunks and narrow it down, if possible.
>
> From the looks of it, I'm thinking:
> 
> * comment out the intel_apply_cpu_quirks() call
> 
> * comment out this hunk
> 
>         /*
>          * There are also broken BIOSes on some Pentium M and
>          * earlier systems:
>          */
>         if (c->x86_vfm < INTEL_CORE_YONAH && mca_cfg.bootlog < 0)
>                 mca_cfg.bootlog = 0;

Making these changes does not appear to have any effect.

I whittled down the revert to the following patch, and this is
sufficient to get the machine booting again:

diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
index 8dd424ac5de8..83eb037d4d46 100644
--- a/arch/x86/kernel/cpu/mce/core.c
+++ b/arch/x86/kernel/cpu/mce/core.c
@@ -2284,6 +2284,12 @@ void mcheck_cpu_init(struct cpuinfo_x86 *c)
 
 	__mcheck_cpu_cap_init();
 
+	{
+	struct mce_bank *mce_banks = this_cpu_ptr(mce_banks_array);
+	if (c->x86_vfm < INTEL_NEHALEM_EP && this_cpu_read(mce_num_banks))
+		mce_banks[0].init = false;
+	}
+
 	if (!mce_gen_pool_init()) {
 		mca_cfg.disabled = 1;
 		pr_emerg("Couldn't allocate MCE records pool!\n");

Re: PROBLEM: boot hang on ASUS P2B-F (regression)

[Yazen Ghannam]

On Sun, Mar 22, 2026 at 11:10:45AM -0400, Nick Bowler wrote:
> On Sat, Mar 21, 2026 at 10:52:24PM +0100, Borislav Petkov wrote:
> > Anyway, since you're partially reverting, can you pls try reverting the
> > Intel-relevant hunks and narrow it down, if possible.
> >
> > From the looks of it, I'm thinking:
> > 
> > * comment out the intel_apply_cpu_quirks() call
> > 
> > * comment out this hunk
> > 
> >         /*
> >          * There are also broken BIOSes on some Pentium M and
> >          * earlier systems:
> >          */
> >         if (c->x86_vfm < INTEL_CORE_YONAH && mca_cfg.bootlog < 0)
> >                 mca_cfg.bootlog = 0;
> 
> Making these changes does not appear to have any effect.
> 
> I whittled down the revert to the following patch, and this is
> sufficient to get the machine booting again:
> 
> diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
> index 8dd424ac5de8..83eb037d4d46 100644
> --- a/arch/x86/kernel/cpu/mce/core.c
> +++ b/arch/x86/kernel/cpu/mce/core.c
> @@ -2284,6 +2284,12 @@ void mcheck_cpu_init(struct cpuinfo_x86 *c)
>  
>  	__mcheck_cpu_cap_init();
>  
> +	{
> +	struct mce_bank *mce_banks = this_cpu_ptr(mce_banks_array);
> +	if (c->x86_vfm < INTEL_NEHALEM_EP && this_cpu_read(mce_num_banks))
> +		mce_banks[0].init = false;
> +	}
> +

Hi Nick,

Can you please test after enabling CONFIG_X86_MCE_INTEL?

Without this, the Intel-specific code, including this quirk, will not be
built.

Thanks,
Yazen