Re: [PATCH 1/1] mm: kfence: skip kmemleak alloc in kfence_pool

[Yee Lee <yee.lee@xxxxxxxxxxxx>]

On Fri, 2022-06-24 at 10:28 +0200, Marco Elver wrote:

On Fri, 24 Jun 2022 at 10:20, 'Yee Lee' via kasan-dev

<kasan-dev@xxxxxxxxxxxxxxxx> wrote:

On Thu, 2022-06-23 at 13:59 +0200, Marco Elver wrote:

On Thu, 23 Jun 2022 at 13:20, yee.lee via kasan-dev

<kasan-dev@xxxxxxxxxxxxxxxx> wrote:

From: Yee Lee <yee.lee@xxxxxxxxxxxx>

Use MEMBLOCK_ALLOC_NOLEAKTRACE to skip kmemleak registration when

the kfence pool is allocated from memblock. And the kmemleak_free

later can be removed too.

Is this purely meant to be a cleanup and non-functional change?

Signed-off-by: Yee Lee <yee.lee@xxxxxxxxxxxx>

---

 mm/kfence/core.c | 18 ++++++++----------

 1 file changed, 8 insertions(+), 10 deletions(-)

diff --git a/mm/kfence/core.c b/mm/kfence/core.c

index 4e7cd4c8e687..0d33d83f5244 100644

--- a/mm/kfence/core.c

+++ b/mm/kfence/core.c

@@ -600,14 +600,6 @@ static unsigned long kfence_init_pool(void)

                addr += 2 * PAGE_SIZE;

        }

-       /*

-        * The pool is live and will never be deallocated from this

point on.

-        * Remove the pool object from the kmemleak object tree, as

it would

-        * otherwise overlap with allocations returned by

kfence_alloc(), which

-        * are registered with kmemleak through the slab post-alloc

hook.

-        */

-       kmemleak_free(__kfence_pool);

This appears to only be a non-functional change if the pool is

allocated early. If the pool is allocated late using page-alloc, then

there'll not be a kmemleak_free() on that memory and we'll have the

same problem.

Do you mean the kzalloc(slab_is_available) in memblock_allc()? That

implies that MEMBLOCK_ALLOC_NOLEAKTRACE has no guarantee skipping

kmemleak_alloc from this. (Maybe add it?)

No, if KFENCE is initialized through kfence_init_late() ->

kfence_init_pool_late() -> kfence_init_pool().

Thanks for the information.

But as I known, page-alloc does not request kmemleak areas.

So the current kfence_pool_init_late() would cause another kmemleak warning on unknown freeing. 

Reproducing test: (kfence late enable + kmemleak debug on)

/ # echo 500 > /sys/module/kfence/parameters/sample_interval

[  153.433518] kmemleak: Freeing unknown object at 0xffff0000c0600000

[  153.433804] CPU: 0 PID: 100 Comm: sh Not tainted 5.19.0-rc3-74069-gde5c208d533a-dirty #1

[  153.434027] Hardware name: linux,dummy-virt (DT)

[  153.434265] Call trace:

[  153.434331]  dump_backtrace+0xdc/0xfc

[  153.434962]  show_stack+0x18/0x24

[  153.435106]  dump_stack_lvl+0x64/0x7c

[  153.435232]  dump_stack+0x18/0x38

[  153.435347]  kmemleak_free+0x184/0x1c8

[  153.435462]  kfence_init_pool+0x16c/0x194

[  153.435587]  param_set_sample_interval+0xe0/0x1c4

[  153.435694]  param_attr_store+0x98/0xf4

[  153.435804]  module_attr_store+0x24/0x3c

[  153.435910]  sysfs_kf_write+0x3c/0x50

...(skip)

[  153.444496] kfence: initialized - using 524288 bytes for 63 objects at 0x00000000a3236b01-0x00000000901655d3

/ # 

Hence, now there are two issues to solve.

(1) (The original)To prevent the undesired kmemleak scanning on the kfence pool. As Cataline's suggestion, we can just apply kmemleak_ignore_phys instead of free it at all. 

ref: https://lore.kernel.org/linux-mm/YrWPg3xIHbm9bFxP@xxxxxxx/

(2) The late-allocated kfence pool doesn't need to go through kmemleak_free. We can relocate the opeartion to kfence_init_pool_early() to seperate them. 

That is, kfence_init_pool_early(memblock) has it and kfence_init_pool_late(page alloc) does not. 

The draft is like the following.

diff --git a/mm/kfence/core.c b/mm/kfence/core.c

index 11a954763be9..a52db7f06c04 100644

--- a/mm/kfence/core.c

+++ b/mm/kfence/core.c

@@ -591,14 +591,6 @@ static unsigned long kfence_init_pool(void)

                addr += 2 * PAGE_SIZE;

        }

-       /*

-        * The pool is live and will never be deallocated from this point on.

-        * Remove the pool object from the kmemleak object tree, as it would

-        * otherwise overlap with allocations returned by kfence_alloc(), which

-        * are registered with kmemleak through the slab post-alloc hook.

-        */

-       kmemleak_free(__kfence_pool);

-

        return 0;

 }

@@ -611,8 +603,16 @@ static bool __init kfence_init_pool_early(void)

        addr = kfence_init_pool();

-       if (!addr)

+       if (!addr) {

+               /*

+                * The pool is live and will never be deallocated from this point on.

+                * Ignore the pool object from the kmemleak phys object tree, as it would

+                * otherwise overlap with allocations returned by kfence_alloc(), which

+                * are registered with kmemleak through the slab post-alloc hook.

+                */

+               kmemleak_ignore_phys(__pa(__kfence_pool));

                return true;

+       }

        /*

         * Only release unprotected pages, and do not try to go back and change