arch/x86/mm/hugetlbpage.c: pud_huge() returns 0 when using 2-level paging
authorNaoya Horiguchi <naoya.horiguchi@nec.com>
Mon, 7 Nov 2022 02:10:10 +0000 (11:10 +0900)
committerAndrew Morton <akpm@linux-foundation.org>
Tue, 8 Nov 2022 23:57:25 +0000 (15:57 -0800)
commit1fdbed657a4726639c4f17841fd2a0fb646c746e
tree4118f39b23a20867f8aa597dc59ac50441fe01ab
parent82e60d00b753bb5cfecce22b8e952436b14d02a3
arch/x86/mm/hugetlbpage.c: pud_huge() returns 0 when using 2-level paging

The following bug is reported to be triggered when starting X on x86-32
system with i915:

  [  225.777375] kernel BUG at mm/memory.c:2664!
  [  225.777391] invalid opcode: 0000 [#1] PREEMPT SMP
  [  225.777405] CPU: 0 PID: 2402 Comm: Xorg Not tainted 6.1.0-rc3-bdg+ #86
  [  225.777415] Hardware name:  /8I865G775-G, BIOS F1 08/29/2006
  [  225.777421] EIP: __apply_to_page_range+0x24d/0x31c
  [  225.777437] Code: ff ff 8b 55 e8 8b 45 cc e8 0a 11 ec ff 89 d8 83 c4 28 5b 5e 5f 5d c3 81 7d e0 a0 ef 96 c1 74 ad 8b 45 d0 e8 2d 83 49 00 eb a3 <0f> 0b 25 00 f0 ff ff 81 eb 00 00 00 40 01 c3 8b 45 ec 8b 00 e8 76
  [  225.777446] EAX: 00000001 EBX: c53a3b58 ECX: b5c00000 EDX: c258aa00
  [  225.777454] ESI: b5c00000 EDI: b5900000 EBP: c4b0fdb4 ESP: c4b0fd80
  [  225.777462] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 EFLAGS: 00010202
  [  225.777470] CR0: 80050033 CR2: b5900000 CR3: 053a3000 CR4: 000006d0
  [  225.777479] Call Trace:
  [  225.777486]  ? i915_memcpy_init_early+0x63/0x63 [i915]
  [  225.777684]  apply_to_page_range+0x21/0x27
  [  225.777694]  ? i915_memcpy_init_early+0x63/0x63 [i915]
  [  225.777870]  remap_io_mapping+0x49/0x75 [i915]
  [  225.778046]  ? i915_memcpy_init_early+0x63/0x63 [i915]
  [  225.778220]  ? mutex_unlock+0xb/0xd
  [  225.778231]  ? i915_vma_pin_fence+0x6d/0xf7 [i915]
  [  225.778420]  vm_fault_gtt+0x2a9/0x8f1 [i915]
  [  225.778644]  ? lock_is_held_type+0x56/0xe7
  [  225.778655]  ? lock_is_held_type+0x7a/0xe7
  [  225.778663]  ? 0xc1000000
  [  225.778670]  __do_fault+0x21/0x6a
  [  225.778679]  handle_mm_fault+0x708/0xb21
  [  225.778686]  ? mt_find+0x21e/0x5ae
  [  225.778696]  exc_page_fault+0x185/0x705
  [  225.778704]  ? doublefault_shim+0x127/0x127
  [  225.778715]  handle_exception+0x130/0x130
  [  225.778723] EIP: 0xb700468a

Recently pud_huge() got aware of non-present entry by commit 3a194f3f8ad0
("mm/hugetlb: make pud_huge() and follow_huge_pud() aware of non-present
pud entry") to handle some special states of gigantic page.  However, it's
overlooked that pud_none() always returns false when running with 2-level
paging, and as a result pud_huge() can return true pointlessly.

Introduce "#if CONFIG_PGTABLE_LEVELS > 2" to pud_huge() to deal with this.

Link: https://lkml.kernel.org/r/20221107021010.2449306-1-naoya.horiguchi@linux.dev
Fixes: 3a194f3f8ad0 ("mm/hugetlb: make pud_huge() and follow_huge_pud() aware of non-present pud entry")
Signed-off-by: Naoya Horiguchi <naoya.horiguchi@nec.com>
Reported-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Tested-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Liu Shixin <liushixin2@huawei.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Muchun Song <songmuchun@bytedance.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Yang Shi <shy828301@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
arch/x86/mm/hugetlbpage.c