linux-kernel - Re: [PATCH] rpmsg: virtio_rpmsg

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <619b06b0-c3cc-e36f-c3e5-d13e98520c9c@ti.com>
Date:   Tue, 27 Aug 2019 17:15:39 -0500
From:   Suman Anna <s-anna@...com>
To:     Bjorn Andersson <bjorn.andersson@...aro.org>
CC:     <linux-remoteproc@...r.kernel.org>,
        Loic Pallardy <loic.pallardy@...com>,
        <linux-kernel@...r.kernel.org>,
        <linux-arm-kernel@...ts.infradead.org>
Subject: Re: [PATCH] rpmsg: virtio_rpmsg_bus: replace "%p" with "%pK"

On 8/27/19 5:07 PM, Bjorn Andersson wrote:
> On Tue 27 Aug 13:25 PDT 2019, Suman Anna wrote:
> 
>> Hi Bjorn,
>>
>> On 8/27/19 12:10 AM, Bjorn Andersson wrote:
>>> On Fri 09 Aug 13:25 PDT 2019, Suman Anna wrote:
>>>
>>>> Hi Bjorn,
>>>>
>>>
>>> Hi Suman
>>>
>>>> On 10/23/18 8:19 PM, Suman Anna wrote:
>>>>> The virtio_rpmsg_bus driver uses the "%p" format-specifier for
>>>>> printing the vring buffer address. This prints only a hashed
>>>>> pointer even for previliged users. Use "%pK" instead so that
>>>>> the address can be printed during debug using kptr_restrict
>>>>> sysctl.
>>>>
>>>> Seems to have been lost among the patches, can you pick up this trivial
>>>> patch for 5.4? Should apply cleanly on the latest HEAD as well.
>>>>
>>>
>>> I share Andrew's question regarding what benefit you have from knowing
>>> this value. Should we not just remove the va from the print? Or do you
>>> actually have a use case for it?.
>>
>> I mainly use it during debug when comparing against kernel_page_tables
>> and vmallocinfo. The pools that we use are not always guaranteed to be
>> from linear memory, and behavior changes when using with CMA or DMA pools.
>>
> 
> Thanks Suman. It seems to me that there's room for improvement to aid
> this kind of debugging. But your usecase seems reasonable, so I'm
> merging the patch.

Thanks Bjorn.

> 
>> Note that usage of %pK does not leak the addresses automatically, but
>> atleast enables me to get the values when needed. The changes also bring
>> the usage in rpmsg core in sync with the remoteproc core.
>>
> 
> Sounds like shouldn't have merged them in remoteproc then ;P

Slightly different reasoning looking at the commit, it was probably when
%p was leaking the addresses.

regards
Suman

> 
> Thanks,
> Bjorn
> 
>> regards
>> Suman
>>
>>>
>>> Regards,
>>> Bjorn
>>>
>>>> regards
>>>> Suman
>>>>
>>>>>
>>>>> Signed-off-by: Suman Anna <s-anna@...com>
>>>>> ---
>>>>>  drivers/rpmsg/virtio_rpmsg_bus.c | 2 +-
>>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/drivers/rpmsg/virtio_rpmsg_bus.c b/drivers/rpmsg/virtio_rpmsg_bus.c
>>>>> index f29dee731026..1345f373a1a0 100644
>>>>> --- a/drivers/rpmsg/virtio_rpmsg_bus.c
>>>>> +++ b/drivers/rpmsg/virtio_rpmsg_bus.c
>>>>> @@ -950,7 +950,7 @@ static int rpmsg_probe(struct virtio_device *vdev)
>>>>>  		goto vqs_del;
>>>>>  	}
>>>>>  
>>>>> -	dev_dbg(&vdev->dev, "buffers: va %p, dma %pad\n",
>>>>> +	dev_dbg(&vdev->dev, "buffers: va %pK, dma %pad\n",
>>>>>  		bufs_va, &vrp->bufs_dma);
>>>>>  
>>>>>  	/* half of the buffers is dedicated for RX */
>>>>>
>>>>
>>