Skip to contentSkip to conversation

arm64: mm: check for upper PAGE_SHIFT bits in pfn_valid()

PATCH New Greg Hackmann 1 participant 2 messages Aug 15, 2018

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/2 arm64: mm: check for upper PAGE_SHIFT bits in pfn_valid()

ARM64's pfn_valid() shifts away the upper PAGE_SHIFT bits of the input before seeing if the PFN is valid. This leads to false positives when some of the upper bits are set, but the lower bits match a valid PFN.

For example, the following userspace code looks up a bogus entry in /proc/kpageflags:

    int pagemap = open("/proc/self/pagemap", O_RDONLY);
    int pageflags = open("/proc/kpageflags", O_RDONLY);
    uint64_t pfn, val;

    lseek64(pagemap, [...], SEEK_SET);
    read(pagemap, &pfn, sizeof(pfn));
    if (pfn & (1UL << 63)) {        /* valid PFN */
        pfn &= ((1UL << 55) - 1);   /* clear flag bits */
        pfn |= (1UL << 55);
        lseek64(pageflags, pfn * sizeof(uint64_t), SEEK_SET);
        read(pageflags, &val, sizeof(val));
    }

On ARM64 this causes the userspace process to crash with SIGSEGV rather than reading (1 << KPF_NOPAGE). kpageflags_read() treats the offset as valid, and stable_page_flags() will try to access an address between the user and kernel address ranges.

Fixes c1cc1552616d ("arm64: MMU initialisation")
Cc stable at vger.kernel.org
---
 arch/arm64/mm/init.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
arch/arm64/mm/init.c +5 -1
index 9abf8a1e7b25..787e27964ab9 100644
@@ -287,7 +287,11 @@ static void __init zone_sizes_init(unsigned long min, unsigned long max)
 #ifdef CONFIG_HAVE_ARCH_PFN_VALID
 int pfn_valid(unsigned long pfn)
 {
-	return memblock_is_map_memory(pfn << PAGE_SHIFT);
+	phys_addr_t addr = pfn << PAGE_SHIFT;
+
+	if ((addr >> PAGE_SHIFT) != pfn)
+		return 0;
+	return memblock_is_map_memory(addr);
 }
 EXPORT_SYMBOL(pfn_valid);
 #endif
Signature · 1 line
2.18.0.865.gffc8e1a3cd6-goog
#1 lore raw

PATCH v2 2/2 arm: mm: check for upper PAGE_SHIFT bits in pfn_valid()

ARM's pfn_valid() has a similar shifting bug to the ARM64 bug fixed in the previous patch. This only affects non-LPAE kernels, since LPAE kernels will promote to 64 bits inside __pfn_to_phys().

Fixes 5e6f6aa1c243 ("memblock/arm: pfn_valid uses memblock_is_memory()")
Cc stable at vger.kernel.org
---
 arch/arm/mm/init.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
arch/arm/mm/init.c +5 -1
index 0cc8e04295a4..bee1f2e4ecf3 100644
@@ -196,7 +196,11 @@ static void __init zone_sizes_init(unsigned long min, unsigned long max_low,
 #ifdef CONFIG_HAVE_ARCH_PFN_VALID
 int pfn_valid(unsigned long pfn)
 {
-	return memblock_is_map_memory(__pfn_to_phys(pfn));
+	phys_addr_t addr = __pfn_to_phys(pfn);
+
+	if (__phys_to_pfn(addr) != pfn)
+		return 0;
+	return memblock_is_map_memory(addr);
 }
 EXPORT_SYMBOL(pfn_valid);
 #endif
Signature · 1 line
2.18.0.865.gffc8e1a3cd6-goog