Hello Daniel,
This test just ensures that PFN is preserved during pte <--> swap pte transformations
, and the warning here seems to have been caused by powerpc platform specific helpers
and/or its pte_t representation. Adding powerpc folks and platform mailing list here.
- Anshuman
On 11/4/23 04:42, Daniel Walker (danielwa) wrote:
>
> Hi,
>
> This warning has triggered on our powerpc equipment.
>
> debug_vm_pgtable: [debug_vm_pgtable ]: Validating architecture page table helpers
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 1 at mm/debug_vm_pgtable.c:860 0xc1012190
> Modules linked in:
> CPU: 0 PID: 1 Comm: swapper Not tainted 6.6.0 #14
> Hardware name: MPC8544DS e500v2 0x80210030 MPC8544 DS
> NIP: c1012190 LR: c101216c CTR: 00000000
> REGS: c145bc70 TRAP: 0700 Not tainted (6.6.0)
> MSR: 00029000 <CE,EE,ME> CR: 48000248 XER: 00000000
>
> GPR00: c1012094 c145bd60 c1498000 c145bd6c 0000003c 00000000 7cffffff 00000001
> GPR08: efa2d000 00000000 00000000 2d589a6c 88000248 00000000 c000343c 00000000
> GPR16: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 c10000d8
> GPR24: 00000000 00000000 00000000 00000000 00000000 00000200 fffffffc 7cffffff
> Call Trace:
> [c145bd60] [c1012094] 0xc1012094 (unreliable)
> [c145be30] [c0003044] 0xc0003044
> [c145bea0] [c1000ea0] 0xc1000ea0
> [c145bee0] [c0003460] 0xc0003460
> [c145bf00] [c0011224] 0xc0011224
> --- interrupt: 0 at 0x00000000
> NIP: 00000000 LR: 00000000 CTR: 00000000
> REGS: c145bf10 TRAP: 0000 Not tainted (6.6.0)
> MSR: 00000000 <> CR: 00000000 XER: 00000000
>
> GPR00: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> GPR08: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> GPR16: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> GPR24: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> --- interrupt: 0
> Code: 3861000c 38a00000 4b10a235 8121000c 7c1f4800 41820008 0fe00000 8121006c 7d29e378 5529463e 7c1d4840 41820008 <0fe00000> 81410080 2c0affff 41820058
> ---[ end trace 0000000000000000 ]---
>
> This appears to be the following code,
>
> 850
> 851 static void __init pte_swap_tests(struct pgtable_debug_args *args)
> 852 {
> 853 swp_entry_t swp;
> 854 pte_t pte;
> 855
> 856 pr_debug("Validating PTE swap\n");
> 857 pte = pfn_pte(args->fixed_pte_pfn, args->page_prot);
> 858 swp = __pte_to_swp_entry(pte);
> 859 pte = __swp_entry_to_pte(swp);
> 860 WARN_ON(args->fixed_pte_pfn != pte_pfn(pte));
> 861 }
> 862
>
>
> This only happens in powerpc with the following config options enabled,
>
> CONFIG_PTE_64BIT=y
> CONFIG_PHYS_64BIT=y
>
> These options can increase the pte_t type to 64bits, which swp_entry_t remains unsigned
> long which is 32 bits on there systems. This warning appears to have triggered since the
> inclusion of CONFIG_DEBUG_VM_PGTABLE support for powerpc.
>
> To reproduce the issue you can use the attached config to build a powerpc kernel. Then
> boot with QEMU using the following command line.
>
> qemu-system-ppc -M mpc8544ds -kernel ../powerpc/vmlinux -append "console=ttyS0" -nographic
>
> Please CC me on any fixes in the future.
>
> Thanks,
>
> Daniel
>
