Re: [PATCH V8] mm/debug: Add tests validating architecture page table helpers

[Michael Ellerman]

Anshuman Khandual <anshuman.khandual@xxxxxxx> writes:
> On 11/06/2019 12:11 PM, Christophe Leroy wrote:
>> Le 06/11/2019 Ã 04:22, Anshuman Khandual a ÃcritÂ:
>>> On 10/28/2019 10:59 AM, Anshuman Khandual wrote:
>>>> +ÂÂÂ -----------------------
>>>> +ÂÂÂ |ÂÂÂÂÂÂÂÂ arch |status|
>>>> +ÂÂÂ -----------------------
>>>> +ÂÂÂ |ÂÂÂÂÂÂ alpha: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂÂÂÂÂ arc: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂÂÂÂÂ arm: | TODO |
>>>> + | arm64: | ok |
>>>> +ÂÂÂ |ÂÂÂÂÂÂÂÂ c6x: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂÂÂÂ csky: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂÂÂ h8300: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂ hexagon: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂÂÂÂ ia64: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂÂÂÂ m68k: | TODO |
>>>> +ÂÂÂ |Â microblaze: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂÂÂÂ mips: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂÂÂ nds32: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂÂÂ nios2: | TODO |
>>>> +ÂÂÂ |ÂÂÂ openrisc: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂÂ parisc: | TODO |
>>>> +ÂÂÂ |ÂÂÂÂ powerpc: | TODO |
>>>> + | ppc32: | ok |
>> 
>> Note that ppc32 is a part of powerpc, not a standalone arch.
>
> Right, I understand. But we are yet to hear about how this test
> came about on powerpc server platforms. Will update 'powerpc'
> arch listing above once we get some confirmation. May be once
> this works on all relevant powerpc platforms, we can just merge
> 'powerpc' and 'ppc32' entries here as just 'powerpc'.

On pseries:

  watchdog: BUG: soft lockup - CPU#0 stuck for 23s! [swapper/0:1]
  Modules linked in:
  CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.4.0-rc6-gcc-8.2.0-next-20191107-00001-g250339d6747b-dirty #152
  NIP:  c0000000010435a0 LR: c0000000010434b4 CTR: 0000000000000000
  REGS: c00000003a403980 TRAP: 0901   Not tainted  (5.4.0-rc6-gcc-8.2.0-next-20191107-00001-g250339d6747b-dirty)
  MSR:  8000000002009033 <SF,VEC,EE,ME,IR,DR,RI,LE>  CR: 44000222  XER: 00000000
  CFAR: c0000000010435a8 IRQMASK: 0 
  GPR00: c0000000010434b4 c00000003a403c10 c000000001295000 05210001000000c0 
  GPR04: 8000000000000105 0000000000400dc0 000000003eb00000 0000000000000001 
  GPR08: 0000000000000000 ffffffffffffffff 0000000000000001 0000000000000100 
  GPR12: 0000000000000000 c0000000018f0000 
  NIP [c0000000010435a0] debug_vm_pgtable+0x43c/0x82c
  LR [c0000000010434b4] debug_vm_pgtable+0x350/0x82c
  Call Trace:
  [c00000003a403c10] [c00000000104346c] debug_vm_pgtable+0x308/0x82c (unreliable)
  [c00000003a403ce0] [c000000001004310] kernel_init_freeable+0x1d0/0x39c
  [c00000003a403db0] [c000000000010da0] kernel_init+0x24/0x174
  [c00000003a403e20] [c00000000000bdc4] ret_from_kernel_thread+0x5c/0x78
  Instruction dump:
  7d075078 7ce74b78 7ce0f9ad 40c2fff0 38800000 7f83e378 4b02eee1 60000000 
  48000080 3920ffff 39400001 39000000 <7ea0f8a8> 7ea75039 40c2fff8 7ea74878 

Looking at the asm I think it's stuck in hash__pte_update() waiting for
H_PAGE_BUSY to clear, but not sure why.

That's just using qemu TCG, instructions here if anyone wants to test it
themselves :)

  https://github.com/linuxppc/wiki/wiki/Booting-with-Qemu


If I boot with -cpu power9 (using Radix MMU), I get a plain old BUG:

  debug_vm_pgtable: debug_vm_pgtable: Validating architecture page table helpers
  ------------[ cut here ]------------
  kernel BUG at arch/powerpc/mm/pgtable.c:274!
  Oops: Exception in kernel mode, sig: 5 [#1]
  LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=32 NUMA pSeries
  Modules linked in:
  CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.4.0-rc6-gcc-8.2.0-next-20191107-00001-g250339d6747b-dirty #152
  NIP:  c0000000000724e8 LR: c00000000104358c CTR: 0000000000000000
  REGS: c00000003a483980 TRAP: 0700   Not tainted  (5.4.0-rc6-gcc-8.2.0-next-20191107-00001-g250339d6747b-dirty)
  MSR:  8000000002029033 <SF,VEC,EE,ME,IR,DR,RI,LE>  CR: 24000224  XER: 20000000
  CFAR: c000000001043588 IRQMASK: 0 
  GPR00: c00000000104358c c00000003a483c10 c000000001295000 0000000000000009 
  GPR04: 0000000000000000 0000000000000005 0000000000000000 0000000000000009 
  GPR08: 0000000000000001 000000000000000e 0000000000000001 c00000003a5f0000 
  GPR12: 0000000000000000 c0000000018f0000 c000000000010d84 0000000000000000 
  GPR16: 0000000000000000 0000000000000000 c00000003a5f0000 8000000000000105 
  GPR20: c000000001003ab8 0000000000000015 0500613a00000080 0900603a00000080 
  GPR24: 09202e3a00000080 c00000000133bd90 c00000000133bd98 c00000000133bda0 
  GPR28: c00000003a5e0000 c00000003a600af8 c00000003a2e2d48 c00000003a6100a0 
  NIP [c0000000000724e8] assert_pte_locked+0x88/0x190
  LR [c00000000104358c] debug_vm_pgtable+0x428/0x82c
  Call Trace:
  [c00000003a483c10] [c00000000104346c] debug_vm_pgtable+0x308/0x82c (unreliable)
  [c00000003a483ce0] [c000000001004310] kernel_init_freeable+0x1d0/0x39c
  [c00000003a483db0] [c000000000010da0] kernel_init+0x24/0x174
  [c00000003a483e20] [c00000000000bdc4] ret_from_kernel_thread+0x5c/0x78
  Instruction dump:
  7d251a14 39070010 7d463030 7d084a14 38c6ffff 7c884436 7cc607b4 7d083038 
  79081f24 7ccb402a 7cc80074 7908d182 <0b080000> 78cb0022 54c8c03e 7d473830 
  ---[ end trace a694f1bc56529c0e ]---


cheers