[PATCH net] net: mlx5: fix sizeof usage in health_care's reg_handler

[PATCH net] net: mlx5: fix sizeof usage in health_care's reg_handler

[Daniel Borkmann]

In mlx5's function health_care() the callback handler reg_handler() is
being called that checks the devices registers like:

  reg_handler(dev->pdev, health->health, sizeof(health->health));

health->health is a pointer to the member "struct health_buffer __iomem
*health" of mlx5_core_health, where health buffer itself looks like:

  struct health_buffer {
        __be32          assert_var[5];
        __be32          rsvd0[3];
        __be32          assert_exit_ptr;
        __be32          assert_callra;
        __be32          rsvd1[2];
        ...
        __be16          ext_sync;
  };

Therefore, I strongly assume sizeof(*health->health) is being meant
to be passed as an argument. Interestingly, there are actually no
in-tree users of mlx5_[un]register_health_report_handler(), but some
debugging modules might want to know the correct size instead.

Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Or Gerlitz <ogerlitz@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/health.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/health.c b/drivers/net/ethernet/mellanox/mlx5/core/health.c
index 748f10a..3592e43 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/health.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/health.c
@@ -101,7 +101,7 @@ static void health_care(struct work_struct *work)
 		spin_lock_irq(&health_lock);
 		if (reg_handler)
 			reg_handler(dev->pdev, health->health,
-				    sizeof(health->health));
+				    sizeof(*health->health));
 
 		list_del_init(&health->list);
 		spin_unlock_irq(&health_lock);
-- 
1.7.11.7

Re: [PATCH net] net: mlx5: fix sizeof usage in health_care's reg_handler

[David Miller]

From: Daniel Borkmann <dborkman@redhat.com>
Date: Fri,  2 Aug 2013 12:16:17 +0200

> Therefore, I strongly assume sizeof(*health->health) is being meant
> to be passed as an argument. Interestingly, there are actually no
> in-tree users of mlx5_[un]register_health_report_handler(), but some
> debugging modules might want to know the correct size instead.

I want these hooks and infrastructure removed immediately.

If there are no in-tree users there is no reason for them to
exist at all.

Thanks.

Re: [PATCH net] net: mlx5: fix sizeof usage in health_care's reg_handler

[Daniel Borkmann]

On 08/03/2013 12:11 AM, David Miller wrote:
> From: Daniel Borkmann <dborkman@redhat.com>
> Date: Fri,  2 Aug 2013 12:16:17 +0200
>
>> Therefore, I strongly assume sizeof(*health->health) is being meant
>> to be passed as an argument. Interestingly, there are actually no
>> in-tree users of mlx5_[un]register_health_report_handler(), but some
>> debugging modules might want to know the correct size instead.
>
> I want these hooks and infrastructure removed immediately.
>
> If there are no in-tree users there is no reason for them to
> exist at all.

Ok, I let Or handle that. Maybe he wants to add a user of it, instead.

I also noticed that coverty scanner found a couple of other issues,
e.g. outlen_write() in mlx5/core/cmd.c does a kzalloc() without
doing sanity checks on the user-passed allocation size, e.g. it
could even be a negative value passed to it.

Re: [PATCH net] net: mlx5: fix sizeof usage in health_care's reg_handler

[Or Gerlitz]

On Sat, Aug 3, 2013 at 1:43 AM, Daniel Borkmann <dborkman@redhat.com> wrote:

> On 08/03/2013 12:11 AM, David Miller wrote:
>>
>> From: Daniel Borkmann <dborkman@redhat.com>
>> Date: Fri,  2 Aug 2013 12:16:17 +0200
>>
>>> Therefore, I strongly assume sizeof(*health->health) is being meant
>>> to be passed as an argument. Interestingly, there are actually no
>>> in-tree users of mlx5_[un]register_health_report_handler(), but some
>>> debugging modules might want to know the correct size instead.
>>
>>
>> I want these hooks and infrastructure removed immediately.
>>
>> If there are no in-tree users there is no reason for them to
>> exist at all.
>
>
> Ok, I let Or handle that. Maybe he wants to add a user of it, instead.
>
> I also noticed that coverty scanner found a couple of other issues,
> e.g. outlen_write() in mlx5/core/cmd.c does a kzalloc() without
> doing sanity checks on the user-passed allocation size, e.g. it
> could even be a negative value passed to it.

Hi Daniel,

As listed in the kernel maintainers file, Eli Cohen (copied) is the
upstream maintainer for the mlx5 driver, please make sure to send/copy him
on patches/questions you have on mlx5. I will let Eli answer/handle your
findings/questions.

Or.

Re: [PATCH net] net: mlx5: fix sizeof usage in health_care's reg_handler

[Eli Cohen]

On Fri, Aug 02, 2013 at 03:11:32PM -0700, David Miller wrote:
> From: Daniel Borkmann <dborkman@redhat.com>
> Date: Fri,  2 Aug 2013 12:16:17 +0200
> 
> > Therefore, I strongly assume sizeof(*health->health) is being meant
> > to be passed as an argument. Interestingly, there are actually no

Daniel, you're fix is correct. I meant to pass the size of the struct
and not the size of the pointer.

> > in-tree users of mlx5_[un]register_health_report_handler(), but some
> > debugging modules might want to know the correct size instead.
> 
> I want these hooks and infrastructure removed immediately.
> 
> If there are no in-tree users there is no reason for them to
> exist at all.
> 

Hi Dave,
the intention here was to allow other module, if they so wish, to get
notified of any problems detected at the device. I understand that you
don't like this but is there another way to handle this requirement
which will be accepted?

Re: [PATCH net] net: mlx5: fix sizeof usage in health_care's reg_handler

[David Miller]

From: Eli Cohen <eli@dev.mellanox.co.il>
Date: Mon, 5 Aug 2013 09:33:34 +0300

> the intention here was to allow other module, if they so wish, to get
> notified of any problems detected at the device. I understand that you
> don't like this but is there another way to handle this requirement
> which will be accepted?

Which modules?  I do not see them.

Please remove this unused facility, now.