linux-kernel - Re: [PATCH] Fix memory leaks in wave5_vpu_open_enc() and wave5_vpu_open

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]

Date:   Tue, 5 Dec 2023 17:36:43 +0000
From:   Robert Beckett <bob.beckett@...labora.com>
To:     Nicolas Dufresne <nicolas.dufresne@...labora.com>,
        Zeng Chi <zengchi@...inos.cn>, nas.chung@...psnmedia.com,
        jackson.lee@...psnmedia.com, mchehab@...nel.org,
        sebastian.fricke@...labora.com, hverkuil-cisco@...all.nl
Cc:     linux-media@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] Fix memory leaks in wave5_vpu_open_enc() and
 wave5_vpu_open_dec()


On 04/12/2023 13:55, Nicolas Dufresne wrote:
> Hi,
>
> Le lundi 04 décembre 2023 à 17:16 +0800, Zeng Chi a écrit :
>> This patch fixes memory leaks on error escapes in wave5_vpu_open_enc()
>> and wave5_vpu_open_dec().
> Please avoid sending twice the same patch. This is still a NAK.

tbf, this is a different patch, concerning the allocation of the 
codec_info within inst, not inst itself.

> regards,
> Nicolas
>
>> Fixes: 9707a6254a8a ("media: chips-media: wave5: Add the v4l2 layer")
>> Signed-off-by: Zeng Chi<zengchi@...inos.cn>
>> ---
>>   drivers/media/platform/chips-media/wave5/wave5-vpu-dec.c | 5 +++--
>>   drivers/media/platform/chips-media/wave5/wave5-vpu-enc.c | 5 +++--
>>   2 files changed, 6 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/media/platform/chips-media/wave5/wave5-vpu-dec.c b/drivers/media/platform/chips-media/wave5/wave5-vpu-dec.c
>> index 8b1417ece96e..b0a045346bb7 100644
>> --- a/drivers/media/platform/chips-media/wave5/wave5-vpu-dec.c
>> +++ b/drivers/media/platform/chips-media/wave5/wave5-vpu-dec.c
>> @@ -1802,9 +1802,10 @@ static int wave5_vpu_open_dec(struct file *filp)
>>   	spin_lock_init(&inst->state_spinlock);
>>   
>>   	inst->codec_info = kzalloc(sizeof(*inst->codec_info), GFP_KERNEL);
>> -	if (!inst->codec_info)
>> +	if (!inst->codec_info) {
>> +		kfree(inst);

for consistency, would be better to jump to cleanup_inst.

Also, maybe consider embedding codec_info  in to struct vpu_instance to 
avoid the double alloc. I've not checked whether this is viable 
throughout the code, but from a quick scan of the original patch, it 
looks like it is always allocated and freed alongside inst.

>>   		return -ENOMEM;
>> -
>> +	}
>>   	v4l2_fh_init(&inst->v4l2_fh, vdev);
>>   	filp->private_data = &inst->v4l2_fh;
>>   	v4l2_fh_add(&inst->v4l2_fh);
>> diff --git a/drivers/media/platform/chips-media/wave5/wave5-vpu-enc.c b/drivers/media/platform/chips-media/wave5/wave5-vpu-enc.c
>> index f29cfa3af94a..bc94de9ea546 100644
>> --- a/drivers/media/platform/chips-media/wave5/wave5-vpu-enc.c
>> +++ b/drivers/media/platform/chips-media/wave5/wave5-vpu-enc.c
>> @@ -1546,9 +1546,10 @@ static int wave5_vpu_open_enc(struct file *filp)
>>   	inst->ops = &wave5_vpu_enc_inst_ops;
>>   
>>   	inst->codec_info = kzalloc(sizeof(*inst->codec_info), GFP_KERNEL);
>> -	if (!inst->codec_info)
>> +	if (!inst->codec_info) {
>> +		kfree(inst);
>>   		return -ENOMEM;
>> -
>> +	}
>>   	v4l2_fh_init(&inst->v4l2_fh, vdev);
>>   	filp->private_data = &inst->v4l2_fh;
>>   	v4l2_fh_add(&inst->v4l2_fh);