The patch titled
Subject: mm/memblock.c: skip kmemleak for kasan_init()
has been added to the -mm tree. Its filename is
mm-memblock-skip-kmemleak-for-kasan_init.patch
This patch should soon appear at
http://ozlabs.org/~akpm/mmots/broken-out/mm-memblock-skip-kmemleak-for-kasan_init.patch
and later at
http://ozlabs.org/~akpm/mmotm/broken-out/mm-memblock-skip-kmemleak-for-kasan_init.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Qian Cai <cai@xxxxxx>
Subject: mm/memblock.c: skip kmemleak for kasan_init()
Kmemleak does not play well with KASAN (tested on both HPE Apollo 70 and
Huawei TaiShan 2280 aarch64 servers).
After calling start_kernel()->setup_arch()->kasan_init(), kmemleak early
log buffer went from something like 280 to 260000 which caused kmemleak
disabled and crash dump memory reservation failed. The multitude of
kmemleak_alloc() calls is from nested loops while KASAN is setting up full
memory mappings, so let early kmemleak allocations skip those
memblock_alloc_internal() calls came from kasan_init() given that those
early KASAN memory mappings should not reference to other memory. Hence,
no kmemleak false positives.
kasan_init
kasan_map_populate [1]
kasan_pgd_populate [2]
kasan_pud_populate [3]
kasan_pmd_populate [4]
kasan_pte_populate [5]
kasan_alloc_zeroed_page
memblock_alloc_try_nid
memblock_alloc_internal
kmemleak_alloc
[1] for_each_memblock(memory, reg)
[2] while (pgdp++, addr = next, addr != end)
[3] while (pudp++, addr = next, addr != end && pud_none(READ_ONCE(*pudp)))
[4] while (pmdp++, addr = next, addr != end && pmd_none(READ_ONCE(*pmdp)))
[5] while (ptep++, addr = next, addr != end && pte_none(READ_ONCE(*ptep)))
Link: http://lkml.kernel.org/r/1543442925-17794-1-git-send-email-cai@xxxxxx
Signed-off-by: Qian Cai <cai@xxxxxx>
Cc: Catalin Marinas <catalin.marinas@xxxxxxx>
Cc: Michal Hocko <mhocko@xxxxxxxx>
Cc: Mike Rapoport <rppt@xxxxxxxxxxxxxxxxxx>
Cc: Alexander Potapenko <glider@xxxxxxxxxx>
Cc: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
arch/arm64/mm/kasan_init.c | 2 +-
include/linux/memblock.h | 1 +
mm/memblock.c | 19 +++++++++++--------
3 files changed, 13 insertions(+), 9 deletions(-)
--- a/arch/arm64/mm/kasan_init.c~mm-memblock-skip-kmemleak-for-kasan_init
+++ a/arch/arm64/mm/kasan_init.c
@@ -39,7 +39,7 @@ static phys_addr_t __init kasan_alloc_ze
{
void *p = memblock_alloc_try_nid(PAGE_SIZE, PAGE_SIZE,
__pa(MAX_DMA_ADDRESS),
- MEMBLOCK_ALLOC_ACCESSIBLE, node);
+ MEMBLOCK_ALLOC_KASAN, node);
return __pa(p);
}
--- a/include/linux/memblock.h~mm-memblock-skip-kmemleak-for-kasan_init
+++ a/include/linux/memblock.h
@@ -357,6 +357,7 @@ static inline int memblock_get_region_no
/* Flags for memblock allocation APIs */
#define MEMBLOCK_ALLOC_ANYWHERE (~(phys_addr_t)0)
#define MEMBLOCK_ALLOC_ACCESSIBLE 0
+#define MEMBLOCK_ALLOC_KASAN 1
/* We are using top down, so it is safe to use 0 here */
#define MEMBLOCK_LOW_LIMIT 0
--- a/mm/memblock.c~mm-memblock-skip-kmemleak-for-kasan_init
+++ a/mm/memblock.c
@@ -262,7 +262,8 @@ phys_addr_t __init_memblock memblock_fin
phys_addr_t kernel_end, ret;
/* pump up @end */
- if (end == MEMBLOCK_ALLOC_ACCESSIBLE)
+ if (end == MEMBLOCK_ALLOC_ACCESSIBLE ||
+ end == MEMBLOCK_ALLOC_KASAN)
end = memblock.current_limit;
/* avoid allocating the first page */
@@ -1482,13 +1483,15 @@ again:
done:
ptr = phys_to_virt(alloc);
- /*
- * The min_count is set to 0 so that bootmem allocated blocks
- * are never reported as leaks. This is because many of these blocks
- * are only referred via the physical address which is not
- * looked up by kmemleak.
- */
- kmemleak_alloc(ptr, size, 0, 0);
+ /* Skip kmemleak for kasan_init() due to high volume. */
+ if (max_addr != MEMBLOCK_ALLOC_KASAN)
+ /*
+ * The min_count is set to 0 so that bootmem allocated
+ * blocks are never reported as leaks. This is because many
+ * of these blocks are only referred via the physical
+ * address which is not looked up by kmemleak.
+ */
+ kmemleak_alloc(ptr, size, 0, 0);
return ptr;
}
_
Patches currently in -mm which might be from cai@xxxxxx are
debugobjects-avoid-recursive-calls-with-kmemleak.patch
debugobjects-call-debug_objects_mem_init-eariler.patch
mm-memblock-skip-kmemleak-for-kasan_init.patch
