Re: [PATCH v3] x86/mm/pat: Fix VM_PAT handling when fork() fails in copy_page_range()

[David Hildenbrand <david@redhat.com>]

On 04.04.25 13:52, Lorenzo Stoakes wrote:
> On Thu, Apr 03, 2025 at 10:59:12PM +0200, David Hildenbrand wrote:
>> On 03.04.25 17:14, Dan Carpenter wrote:
>>> Sorry, I've been having trouble with my email recently...  I replied
>>> earlier but my email got eaten on the way out.
>>>
>>> What happened here is that the zero day bot emails go to me first and
>>> then I review them or forward them depending on if they're a real
>>> issue or not.
>>>
>>> Here it's a false postive because it's set and used if the
>>> (src_vma->vm_flags & VM_PFNMAP) flag is set.  Smatch doesn't parse
>>> this correctly.  I've been meaning to fix this in Smatch for a
>>> while.
>>
>> There is a slight complication (on top of the VM_PFNMAP checks):
>>
>> If "src_vma->vm_flags & VM_PAT" we
>> * set pfn
>> * set dst_vma->vm_flags |= VM_PFNMAP
>>
>> Then, we only consume the pfn if "dst_vma->vm_flags & VM_PFNMAP"
>>
>> While we won't be using the uninitialized pfn (good), we'd still pass an
>> uninitialized pfn, which IIRC is UB; likely nothing happens on GCC clang,
>> but we better handle it.
>>
>> So that should better be changed; I'll send a fix.
> 
> Maybe just worth setting pfn = 0 _as well_ in the caller, belts + braces maybe?

I'm planning on doing the following, just didn't get to testing it:


 From d340fac886c4a15d39d8e963aa8c647b19589413 Mon Sep 17 00:00:00 2001
From: David Hildenbrand <david@redhat.com>
Date: Tue, 29 Oct 2024 22:03:31 +0100
Subject: [PATCH] x86/mm/pat: (un)track_pfn_copy() fix + improvements

We got a late smatch warning and some additional review feedback.

	smatch warnings:
	mm/memory.c:1428 copy_page_range() error: uninitialized symbol 'pfn'.

We actually use the pfn only when it is properly initialized; however,
we may pass an uninitialized value to a function -- although it will not
use it that likely still is UB in C.

Fix it by always initializing pfn when track_pfn_copy() returns 0 --
just as documented.

While at it, clarify the doc of untrack_pfn_copy(), that internal checks
make sure if we actually have to untrack anything.

Fixes: dc84bc2aba85 ("x86/mm/pat: Fix VM_PAT handling when fork() fails in copy_page_range()")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <error27@gmail.com>
Closes: https://lore.kernel.org/r/202503270941.IFILyNCX-lkp@intel.com/
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@surriel.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Hildenbrand <david@redhat.com>
---
  arch/x86/mm/pat/memtype.c | 4 +++-
  include/linux/pgtable.h   | 5 ++++-
  2 files changed, 7 insertions(+), 2 deletions(-)

diff --git a/arch/x86/mm/pat/memtype.c b/arch/x86/mm/pat/memtype.c
index 72d8cbc611583..9ad3e5b055d8a 100644
--- a/arch/x86/mm/pat/memtype.c
+++ b/arch/x86/mm/pat/memtype.c
@@ -992,8 +992,10 @@ int track_pfn_copy(struct vm_area_struct *dst_vma,
  	pgprot_t pgprot;
  	int rc;
  
-	if (!(src_vma->vm_flags & VM_PAT))
+	if (!(src_vma->vm_flags & VM_PAT)) {
+		*pfn = 0;
  		return 0;
+	}
  
  	/*
  	 * Duplicate the PAT information for the dst VMA based on the src
diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h
index e2b705c149454..9457064292141 100644
--- a/include/linux/pgtable.h
+++ b/include/linux/pgtable.h
@@ -1517,12 +1517,15 @@ static inline void track_pfn_insert(struct vm_area_struct *vma, pgprot_t *prot,
  static inline int track_pfn_copy(struct vm_area_struct *dst_vma,
  		struct vm_area_struct *src_vma, unsigned long *pfn)
  {
+	*pfn = 0;
  	return 0;
  }
  
  /*
   * untrack_pfn_copy is called when a VM_PFNMAP VMA failed to copy during
- * copy_page_range(), but after track_pfn_copy() was already called.
+ * copy_page_range(), but after track_pfn_copy() was already called. Can
+ * be called even if track_pfn_copy() did not actually track anything:
+ * handled internally.
   */
  static inline void untrack_pfn_copy(struct vm_area_struct *dst_vma,
  		unsigned long pfn)
-- 
2.48.1


> 
> I mean the patch is already upstream at dc84bc2aba85. So I guess these fixes are
> intended for rc generally?

Yes.

-- 
Cheers,

David / dhildenb