[PATCH] mm: kmemleak: Ignore kmemleak scanning on CMA regions

[PATCH] mm: kmemleak: Ignore kmemleak scanning on CMA regions

[Calvin Zhang]

Just like this:
commit 620951e27457 ("mm/cma: make kmemleak ignore CMA regions").

Add kmemleak_ignore_phys() for CMA created from of reserved node.

Signed-off-by: Calvin Zhang <calvinzhang.cool@gmail.com>
---
 kernel/dma/contiguous.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/kernel/dma/contiguous.c b/kernel/dma/contiguous.c
index 3d63d91cba5c..66bd9a59615e 100644
--- a/kernel/dma/contiguous.c
+++ b/kernel/dma/contiguous.c
@@ -50,6 +50,7 @@
 #include <linux/sizes.h>
 #include <linux/dma-map-ops.h>
 #include <linux/cma.h>
+#include <linux/kmemleak.h>
 
 #ifdef CONFIG_CMA_SIZE_MBYTES
 #define CMA_SIZE_MBYTES CONFIG_CMA_SIZE_MBYTES
@@ -426,6 +427,9 @@ static int __init rmem_cma_setup(struct reserved_mem *rmem)
 		pr_err("Reserved memory: unable to setup CMA region\n");
 		return err;
 	}
+
+	kmemleak_ignore_phys(rmem->base);
+
 	/* Architecture specific contiguous memory fixup. */
 	dma_contiguous_early_fixup(rmem->base, rmem->size);
 
-- 
2.30.2

Re: [PATCH] mm: kmemleak: Ignore kmemleak scanning on CMA regions

[Andrew Morton]

On Fri, 26 Nov 2021 10:47:11 +0800 Calvin Zhang <calvinzhang.cool@gmail.com> wrote:

> Just like this:
> commit 620951e27457 ("mm/cma: make kmemleak ignore CMA regions").
> 
> Add kmemleak_ignore_phys() for CMA created from of reserved node.

Could we please have a full, standalone changelog for this patch?

The 620951e27457 changelog says "Without this, the kernel crashes...". 
Does your patch also fix a crash?  If so under what circumstances and
should we backport this fix into -stable kernels?

Etcetera.

Re: [PATCH] mm: kmemleak: Ignore kmemleak scanning on CMA regions

[Calvin Zhang]

On Sat, Nov 27, 2021 at 04:07:18PM -0800, Andrew Morton wrote:
>On Fri, 26 Nov 2021 10:47:11 +0800 Calvin Zhang <calvinzhang.cool@gmail.com> wrote:
>
>> Just like this:
>> commit 620951e27457 ("mm/cma: make kmemleak ignore CMA regions").
>> 
>> Add kmemleak_ignore_phys() for CMA created from of reserved node.
>
>Could we please have a full, standalone changelog for this patch?

Yes, I will add.

>
>The 620951e27457 changelog says "Without this, the kernel crashes...". 
>Does your patch also fix a crash?  If so under what circumstances and
>should we backport this fix into -stable kernels?

No crash occurred. 620951e27457 avoids crashes caused by accessing
highmem and it was fixed later. Now kmemleak_alloc_phys() and
kmemleak_ignore_phys() skip highmem. This patch is based on the
point that CMA regions don't contain pointers to other kmemleak
objects, and ignores CMA regions from reserved memory as what
620951e27457 did.

--
Thanks
Calvin

Re: [PATCH] mm: kmemleak: Ignore kmemleak scanning on CMA regions

[Catalin Marinas]

On Sun, Nov 28, 2021 at 09:50:53AM +0800, Calvin Zhang wrote:
> On Sat, Nov 27, 2021 at 04:07:18PM -0800, Andrew Morton wrote:
> >On Fri, 26 Nov 2021 10:47:11 +0800 Calvin Zhang <calvinzhang.cool@gmail.com> wrote:
> >> Just like this:
> >> commit 620951e27457 ("mm/cma: make kmemleak ignore CMA regions").
> >> 
> >> Add kmemleak_ignore_phys() for CMA created from of reserved node.
[...]
> >The 620951e27457 changelog says "Without this, the kernel crashes...". 
> >Does your patch also fix a crash?  If so under what circumstances and
> >should we backport this fix into -stable kernels?
> 
> No crash occurred. 620951e27457 avoids crashes caused by accessing
> highmem and it was fixed later. Now kmemleak_alloc_phys() and
> kmemleak_ignore_phys() skip highmem. This patch is based on the
> point that CMA regions don't contain pointers to other kmemleak
> objects, and ignores CMA regions from reserved memory as what
> 620951e27457 did.

Note that kmemleak_ignore() only works if there was a prior
kmemleak_alloc() on that address range. With the previous commit we get
this via the memblock_alloc_range() but I fail to see one on the
rmem_cma_setup() path.

-- 
Catalin

Re: [PATCH] mm: kmemleak: Ignore kmemleak scanning on CMA regions

[Calvin Zhang]

On Thu, Dec 02, 2021 at 06:11:12PM +0000, Catalin Marinas wrote:
>On Sun, Nov 28, 2021 at 09:50:53AM +0800, Calvin Zhang wrote:
>> On Sat, Nov 27, 2021 at 04:07:18PM -0800, Andrew Morton wrote:
>> >On Fri, 26 Nov 2021 10:47:11 +0800 Calvin Zhang <calvinzhang.cool@gmail.com> wrote:
>> >> Just like this:
>> >> commit 620951e27457 ("mm/cma: make kmemleak ignore CMA regions").
>> >> 
>> >> Add kmemleak_ignore_phys() for CMA created from of reserved node.
>[...]
>> >The 620951e27457 changelog says "Without this, the kernel crashes...". 
>> >Does your patch also fix a crash?  If so under what circumstances and
>> >should we backport this fix into -stable kernels?
>> 
>> No crash occurred. 620951e27457 avoids crashes caused by accessing
>> highmem and it was fixed later. Now kmemleak_alloc_phys() and
>> kmemleak_ignore_phys() skip highmem. This patch is based on the
>> point that CMA regions don't contain pointers to other kmemleak
>> objects, and ignores CMA regions from reserved memory as what
>> 620951e27457 did.
>
>Note that kmemleak_ignore() only works if there was a prior
>kmemleak_alloc() on that address range. With the previous commit we get
>this via the memblock_alloc_range() but I fail to see one on the
>rmem_cma_setup() path.

rmem is from memblock_reserve() or early_init_dt_alloc_reserved_memory_arch()
kmemleak_alloc() is not called in the first case. And It's bad to add one.

I think all the reserved regions should be allocated from memblock without
kmemleak_alloc() and let rmem handler choose to add it as kmemleak object
by kmemleak_alloc(). Because MEMBLOCK_ALLOC_NOLEAKTRACE conflicts with range
parameter in memlbock_alloc_* series, all reserved regions and default CMA
region are allocated with kmemleak_alloc().

I think it's better to add memblock_alloc_* series a spearate flag paramter
(like "NOLEAKTRACE") instead of encoding MEMBLOCK_ALLOC_NOLEAKTRACE in `end`
parameter.

--
Calvin