2019-01-23 10:37:25

by Nicholas Mc Guire

[permalink] [raw]
Subject: [PATCH] livepatch: fix size mismatch

kzalloc(sizeof(int)) is called for an int object but then
passed into klp_shadow_alloc() using the size of the pointer.
This probably is not a problem as it will fit - but it should
be cleaned (after all this is reference code).

Signed-off-by: Nicholas Mc Guire <[email protected]>
---

Reported by coccicheck
samples/livepatch//livepatch-shadow-fix1.c:97:30-36: ERROR: application of sizeof to pointer

Patch was compile tested with: x86_64_defconfig + FTRACE=y,
FUNCTION_TRACER=y, SAMPLES=y, LIVEPATCH=y SAMPLE_LIVEPATCH=m

Patch is against 5.0-rc3 (localversion-next is next-20190123)

samples/livepatch/livepatch-shadow-fix1.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/samples/livepatch/livepatch-shadow-fix1.c b/samples/livepatch/livepatch-shadow-fix1.c
index a5a5cac..643ffd5 100644
--- a/samples/livepatch/livepatch-shadow-fix1.c
+++ b/samples/livepatch/livepatch-shadow-fix1.c
@@ -94,7 +94,7 @@ struct dummy *livepatch_fix1_dummy_alloc(void)
return NULL;
}

- klp_shadow_alloc(d, SV_LEAK, sizeof(leak), GFP_KERNEL,
+ klp_shadow_alloc(d, SV_LEAK, sizeof(*leak), GFP_KERNEL,
shadow_leak_ctor, leak);

pr_info("%s: dummy @ %p, expires @ %lx\n",
--
2.1.4



2019-01-23 15:01:03

by Miroslav Benes

[permalink] [raw]
Subject: Re: [PATCH] livepatch: fix size mismatch

On Wed, 23 Jan 2019, Nicholas Mc Guire wrote:

> kzalloc(sizeof(int)) is called for an int object but then
> passed into klp_shadow_alloc() using the size of the pointer.
> This probably is not a problem as it will fit - but it should
> be cleaned (after all this is reference code).
>
> Signed-off-by: Nicholas Mc Guire <[email protected]>
> ---
>
> Reported by coccicheck
> samples/livepatch//livepatch-shadow-fix1.c:97:30-36: ERROR: application of sizeof to pointer
>
> Patch was compile tested with: x86_64_defconfig + FTRACE=y,
> FUNCTION_TRACER=y, SAMPLES=y, LIVEPATCH=y SAMPLE_LIVEPATCH=m
>
> Patch is against 5.0-rc3 (localversion-next is next-20190123)
>
> samples/livepatch/livepatch-shadow-fix1.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/samples/livepatch/livepatch-shadow-fix1.c b/samples/livepatch/livepatch-shadow-fix1.c
> index a5a5cac..643ffd5 100644
> --- a/samples/livepatch/livepatch-shadow-fix1.c
> +++ b/samples/livepatch/livepatch-shadow-fix1.c
> @@ -94,7 +94,7 @@ struct dummy *livepatch_fix1_dummy_alloc(void)
> return NULL;
> }
>
> - klp_shadow_alloc(d, SV_LEAK, sizeof(leak), GFP_KERNEL,
> + klp_shadow_alloc(d, SV_LEAK, sizeof(*leak), GFP_KERNEL,
> shadow_leak_ctor, leak);

I think it is actually fine. We allocate something sizeof(int), but it is
not important much. The leaked pointer is important. We attach the pointer
as a shadow variable, so we can free it later. Thus, the size of the
pointer is important.

Moreover, you're introducing this with the change:

samples/livepatch/livepatch-shadow-fix1.c:97:38: warning: expression using sizeof(void)

Miroslav