diff mbox series

[v2] vhost: fix build

Message ID 20220829082915.1056865-1-zhoumin@loongson.cn (mailing list archive)
State Rejected, archived
Delegated to: Maxime Coquelin
Headers
Series [v2] vhost: fix build |

Checks

Context Check Description
ci/checkpatch warning coding style issues
ci/iol-mellanox-Performance success Performance Testing PASS
ci/iol-aarch64-unit-testing success Testing PASS
ci/iol-intel-Performance success Performance Testing PASS
ci/iol-intel-Functional success Functional Testing PASS
ci/iol-x86_64-unit-testing success Testing PASS
ci/iol-aarch64-compile-testing success Testing PASS
ci/iol-x86_64-compile-testing success Testing PASS
ci/github-robot: build success github build: passed

Commit Message

zhoumin Aug. 29, 2022, 8:29 a.m. UTC 
  On CentOS 8 or Debian 10.4 systems using gcc 12.1 to cross
compile DPDK, gcc shows a following warning which will cause
build to fail when build is run with -werror:

In function 'mbuf_to_desc',
    inlined from 'vhost_enqueue_async_packed' at ../lib/vhost/virtio_net.c:1826:6,
    inlined from 'virtio_dev_rx_async_packed' at ../lib/vhost/virtio_net.c:1840:6,
    inlined from 'virtio_dev_rx_async_submit_packed.constprop' at ../lib/vhost/virtio_net.c:1900:7:
../lib/vhost/virtio_net.c:1161:35: error: 'buf_vec[0].buf_len' may be used uninitialized [-Werror=maybe-uninitialized]
 1161 |         buf_len = buf_vec[vec_idx].buf_len;
      |                   ~~~~~~~~~~~~~~~~^~~~~~~~
../lib/vhost/virtio_net.c: In function 'virtio_dev_rx_async_submit_packed.constprop':
../lib/vhost/virtio_net.c:1838:27: note: 'buf_vec' declared here
 1838 |         struct buf_vector buf_vec[BUF_VECTOR_MAX];
      |                           ^~~~~~~
cc1: all warnings being treated as errors

Actually, there are eight places to see the same codes in the file
lib/vhost/virtio_net.c, and all these `buf_vec` arraies are
initialized by sub-function calls under various conditions.

Although It's hard to understand why gcc just emits warning at one
of the eight places, adding validity checks for array length is
reasonable and can also fix the warning.

Signed-off-by: David Marchand <david.marchand@redhat.com>
Signed-off-by: Min Zhou <zhoumin@loongson.cn>
---
 lib/vhost/virtio_net.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

Comments

zhoumin Sept. 1, 2022, 12:06 a.m. UTC | #1 
Sincerely ping.


On 2022年08月29日 16:29, Min Zhou wrote:
> On CentOS 8 or Debian 10.4 systems using gcc 12.1 to cross
> compile DPDK, gcc shows a following warning which will cause
> build to fail when build is run with -werror:
>
> In function 'mbuf_to_desc',
>      inlined from 'vhost_enqueue_async_packed' at ../lib/vhost/virtio_net.c:1826:6,
>      inlined from 'virtio_dev_rx_async_packed' at ../lib/vhost/virtio_net.c:1840:6,
>      inlined from 'virtio_dev_rx_async_submit_packed.constprop' at ../lib/vhost/virtio_net.c:1900:7:
> ../lib/vhost/virtio_net.c:1161:35: error: 'buf_vec[0].buf_len' may be used uninitialized [-Werror=maybe-uninitialized]
>   1161 |         buf_len = buf_vec[vec_idx].buf_len;
>        |                   ~~~~~~~~~~~~~~~~^~~~~~~~
> ../lib/vhost/virtio_net.c: In function 'virtio_dev_rx_async_submit_packed.constprop':
> ../lib/vhost/virtio_net.c:1838:27: note: 'buf_vec' declared here
>   1838 |         struct buf_vector buf_vec[BUF_VECTOR_MAX];
>        |                           ^~~~~~~
> cc1: all warnings being treated as errors
>
> Actually, there are eight places to see the same codes in the file
> lib/vhost/virtio_net.c, and all these `buf_vec` arraies are
> initialized by sub-function calls under various conditions.
>
> Although It's hard to understand why gcc just emits warning at one
> of the eight places, adding validity checks for array length is
> reasonable and can also fix the warning.
>
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> Signed-off-by: Min Zhou <zhoumin@loongson.cn>
> ---
>   lib/vhost/virtio_net.c | 5 ++++-
>   1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
> index 35fa4670fd..99233f1759 100644
> --- a/lib/vhost/virtio_net.c
> +++ b/lib/vhost/virtio_net.c
> @@ -1153,7 +1153,7 @@ mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq,
>   	struct virtio_net_hdr_mrg_rxbuf tmp_hdr, *hdr = NULL;
>   	struct vhost_async *async = vq->async;
>   
> -	if (unlikely(m == NULL))
> +	if (unlikely(m == NULL || nr_vec == 0))
>   		return -1;
>   
>   	buf_addr = buf_vec[vec_idx].buf_addr;
> @@ -2673,6 +2673,9 @@ desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq,
>   	struct vhost_async *async = vq->async;
>   	struct async_inflight_info *pkts_info;
>   
> +	if (unlikely(nr_vec == 0))
> +		return -1;
> +
>   	buf_addr = buf_vec[vec_idx].buf_addr;
>   	buf_iova = buf_vec[vec_idx].buf_iova;
>   	buf_len = buf_vec[vec_idx].buf_len;

--
Thanks,
Min Zhou
zhoumin Sept. 9, 2022, 9:07 a.m. UTC | #2 
Kindly ping.

Any feedback is appreciated.


On 2022年08月29日 16:29, Min Zhou wrote:
> On CentOS 8 or Debian 10.4 systems using gcc 12.1 to cross
> compile DPDK, gcc shows a following warning which will cause
> build to fail when build is run with -werror:
>
> In function 'mbuf_to_desc',
>      inlined from 'vhost_enqueue_async_packed' at ../lib/vhost/virtio_net.c:1826:6,
>      inlined from 'virtio_dev_rx_async_packed' at ../lib/vhost/virtio_net.c:1840:6,
>      inlined from 'virtio_dev_rx_async_submit_packed.constprop' at ../lib/vhost/virtio_net.c:1900:7:
> ../lib/vhost/virtio_net.c:1161:35: error: 'buf_vec[0].buf_len' may be used uninitialized [-Werror=maybe-uninitialized]
>   1161 |         buf_len = buf_vec[vec_idx].buf_len;
>        |                   ~~~~~~~~~~~~~~~~^~~~~~~~
> ../lib/vhost/virtio_net.c: In function 'virtio_dev_rx_async_submit_packed.constprop':
> ../lib/vhost/virtio_net.c:1838:27: note: 'buf_vec' declared here
>   1838 |         struct buf_vector buf_vec[BUF_VECTOR_MAX];
>        |                           ^~~~~~~
> cc1: all warnings being treated as errors
>
> Actually, there are eight places to see the same codes in the file
> lib/vhost/virtio_net.c, and all these `buf_vec` arraies are
> initialized by sub-function calls under various conditions.
>
> Although It's hard to understand why gcc just emits warning at one
> of the eight places, adding validity checks for array length is
> reasonable and can also fix the warning.
>
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> Signed-off-by: Min Zhou <zhoumin@loongson.cn>
> ---
>   lib/vhost/virtio_net.c | 5 ++++-
>   1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
> index 35fa4670fd..99233f1759 100644
> --- a/lib/vhost/virtio_net.c
> +++ b/lib/vhost/virtio_net.c
> @@ -1153,7 +1153,7 @@ mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq,
>   	struct virtio_net_hdr_mrg_rxbuf tmp_hdr, *hdr = NULL;
>   	struct vhost_async *async = vq->async;
>   
> -	if (unlikely(m == NULL))
> +	if (unlikely(m == NULL || nr_vec == 0))
>   		return -1;
>   
>   	buf_addr = buf_vec[vec_idx].buf_addr;
> @@ -2673,6 +2673,9 @@ desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq,
>   	struct vhost_async *async = vq->async;
>   	struct async_inflight_info *pkts_info;
>   
> +	if (unlikely(nr_vec == 0))
> +		return -1;
> +
>   	buf_addr = buf_vec[vec_idx].buf_addr;
>   	buf_iova = buf_vec[vec_idx].buf_iova;
>   	buf_len = buf_vec[vec_idx].buf_len;
Chenbo Xia Sept. 26, 2022, 2:57 a.m. UTC | #3 
Hi Min,

> -----Original Message-----
> From: Min Zhou <zhoumin@loongson.cn>
> Sent: Monday, August 29, 2022 4:29 PM
> To: david.marchand@redhat.com; maxime.coquelin@redhat.com; Xia, Chenbo
> <chenbo.xia@intel.com>; zhoumin@loongson.cn
> Cc: dev@dpdk.org; maobibo@loongson.cn
> Subject: [PATCH v2] vhost: fix build
> 
> On CentOS 8 or Debian 10.4 systems using gcc 12.1 to cross
> compile DPDK, gcc shows a following warning which will cause
> build to fail when build is run with -werror:
> 
> In function 'mbuf_to_desc',
>     inlined from 'vhost_enqueue_async_packed'
> at ../lib/vhost/virtio_net.c:1826:6,
>     inlined from 'virtio_dev_rx_async_packed'
> at ../lib/vhost/virtio_net.c:1840:6,
>     inlined from 'virtio_dev_rx_async_submit_packed.constprop'
> at ../lib/vhost/virtio_net.c:1900:7:
> ../lib/vhost/virtio_net.c:1161:35: error: 'buf_vec[0].buf_len' may be used
> uninitialized [-Werror=maybe-uninitialized]
>  1161 |         buf_len = buf_vec[vec_idx].buf_len;
>       |                   ~~~~~~~~~~~~~~~~^~~~~~~~
> ../lib/vhost/virtio_net.c: In function
> 'virtio_dev_rx_async_submit_packed.constprop':
> ../lib/vhost/virtio_net.c:1838:27: note: 'buf_vec' declared here
>  1838 |         struct buf_vector buf_vec[BUF_VECTOR_MAX];
>       |                           ^~~~~~~
> cc1: all warnings being treated as errors
> 
> Actually, there are eight places to see the same codes in the file
> lib/vhost/virtio_net.c, and all these `buf_vec` arraies are
> initialized by sub-function calls under various conditions.
> 
> Although It's hard to understand why gcc just emits warning at one
> of the eight places, adding validity checks for array length is
> reasonable and can also fix the warning.
> 
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> Signed-off-by: Min Zhou <zhoumin@loongson.cn>
> ---
>  lib/vhost/virtio_net.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)

Just want you to know that your patch is still pending because by accident
your fix is almost the same as a previous patch that fixes a real issue but
that patch is still in progress:

http://patchwork.dpdk.org/project/dpdk/patch/20220802004938.23670-2-cfontana@suse.de/

Thanks,
Chenbo

> 
> diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
> index 35fa4670fd..99233f1759 100644
> --- a/lib/vhost/virtio_net.c
> +++ b/lib/vhost/virtio_net.c
> @@ -1153,7 +1153,7 @@ mbuf_to_desc(struct virtio_net *dev, struct
> vhost_virtqueue *vq,
>  	struct virtio_net_hdr_mrg_rxbuf tmp_hdr, *hdr = NULL;
>  	struct vhost_async *async = vq->async;
> 
> -	if (unlikely(m == NULL))
> +	if (unlikely(m == NULL || nr_vec == 0))
>  		return -1;
> 
>  	buf_addr = buf_vec[vec_idx].buf_addr;
> @@ -2673,6 +2673,9 @@ desc_to_mbuf(struct virtio_net *dev, struct
> vhost_virtqueue *vq,
>  	struct vhost_async *async = vq->async;
>  	struct async_inflight_info *pkts_info;
> 
> +	if (unlikely(nr_vec == 0))
> +		return -1;
> +
>  	buf_addr = buf_vec[vec_idx].buf_addr;
>  	buf_iova = buf_vec[vec_idx].buf_iova;
>  	buf_len = buf_vec[vec_idx].buf_len;
> --
> 2.31.1
zhoumin Sept. 26, 2022, 3:25 a.m. UTC | #4 
Hi Chenbo,


On Mon, 26 Sep 2022, 10:57, Xia, Chenbo wrote:
> Hi Min,
>
>> -----Original Message-----
>> From: Min Zhou <zhoumin@loongson.cn>
>> Sent: Monday, August 29, 2022 4:29 PM
>> To: david.marchand@redhat.com; maxime.coquelin@redhat.com; Xia, Chenbo
>> <chenbo.xia@intel.com>; zhoumin@loongson.cn
>> Cc: dev@dpdk.org; maobibo@loongson.cn
>> Subject: [PATCH v2] vhost: fix build
>>
>> On CentOS 8 or Debian 10.4 systems using gcc 12.1 to cross
>> compile DPDK, gcc shows a following warning which will cause
>> build to fail when build is run with -werror:
>>
>> In function 'mbuf_to_desc',
>>      inlined from 'vhost_enqueue_async_packed'
>> at ../lib/vhost/virtio_net.c:1826:6,
>>      inlined from 'virtio_dev_rx_async_packed'
>> at ../lib/vhost/virtio_net.c:1840:6,
>>      inlined from 'virtio_dev_rx_async_submit_packed.constprop'
>> at ../lib/vhost/virtio_net.c:1900:7:
>> ../lib/vhost/virtio_net.c:1161:35: error: 'buf_vec[0].buf_len' may be used
>> uninitialized [-Werror=maybe-uninitialized]
>>   1161 |         buf_len = buf_vec[vec_idx].buf_len;
>>        |                   ~~~~~~~~~~~~~~~~^~~~~~~~
>> ../lib/vhost/virtio_net.c: In function
>> 'virtio_dev_rx_async_submit_packed.constprop':
>> ../lib/vhost/virtio_net.c:1838:27: note: 'buf_vec' declared here
>>   1838 |         struct buf_vector buf_vec[BUF_VECTOR_MAX];
>>        |                           ^~~~~~~
>> cc1: all warnings being treated as errors
>>
>> Actually, there are eight places to see the same codes in the file
>> lib/vhost/virtio_net.c, and all these `buf_vec` arraies are
>> initialized by sub-function calls under various conditions.
>>
>> Although It's hard to understand why gcc just emits warning at one
>> of the eight places, adding validity checks for array length is
>> reasonable and can also fix the warning.
>>
>> Signed-off-by: David Marchand <david.marchand@redhat.com>
>> Signed-off-by: Min Zhou <zhoumin@loongson.cn>
>> ---
>>   lib/vhost/virtio_net.c | 5 ++++-
>>   1 file changed, 4 insertions(+), 1 deletion(-)
> Just want you to know that your patch is still pending because by accident
> your fix is almost the same as a previous patch that fixes a real issue but
> that patch is still in progress:
>
> http://patchwork.dpdk.org/project/dpdk/patch/20220802004938.23670-2-cfontana@suse.de/
>
> Thanks,
> Chenbo

Thanks for your helpful reply.
I think I can drop this patch if the patch you mentioned above could be 
accepted.

Thanks,
Min Zhou
>> diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
>> index 35fa4670fd..99233f1759 100644
>> --- a/lib/vhost/virtio_net.c
>> +++ b/lib/vhost/virtio_net.c
>> @@ -1153,7 +1153,7 @@ mbuf_to_desc(struct virtio_net *dev, struct
>> vhost_virtqueue *vq,
>>   	struct virtio_net_hdr_mrg_rxbuf tmp_hdr, *hdr = NULL;
>>   	struct vhost_async *async = vq->async;
>>
>> -	if (unlikely(m == NULL))
>> +	if (unlikely(m == NULL || nr_vec == 0))
>>   		return -1;
>>
>>   	buf_addr = buf_vec[vec_idx].buf_addr;
>> @@ -2673,6 +2673,9 @@ desc_to_mbuf(struct virtio_net *dev, struct
>> vhost_virtqueue *vq,
>>   	struct vhost_async *async = vq->async;
>>   	struct async_inflight_info *pkts_info;
>>
>> +	if (unlikely(nr_vec == 0))
>> +		return -1;
>> +
>>   	buf_addr = buf_vec[vec_idx].buf_addr;
>>   	buf_iova = buf_vec[vec_idx].buf_iova;
>>   	buf_len = buf_vec[vec_idx].buf_len;
>> --
>> 2.31.1
Maxime Coquelin Oct. 5, 2022, 2:51 p.m. UTC | #5 
Hi Min,

On 9/26/22 05:25, zhoumin wrote:
> Hi Chenbo,
> 
> 
> On Mon, 26 Sep 2022, 10:57, Xia, Chenbo wrote:
>> Hi Min,
>>
>>> -----Original Message-----
>>> From: Min Zhou <zhoumin@loongson.cn>
>>> Sent: Monday, August 29, 2022 4:29 PM
>>> To: david.marchand@redhat.com; maxime.coquelin@redhat.com; Xia, Chenbo
>>> <chenbo.xia@intel.com>; zhoumin@loongson.cn
>>> Cc: dev@dpdk.org; maobibo@loongson.cn
>>> Subject: [PATCH v2] vhost: fix build
>>>
>>> On CentOS 8 or Debian 10.4 systems using gcc 12.1 to cross
>>> compile DPDK, gcc shows a following warning which will cause
>>> build to fail when build is run with -werror:
>>>
>>> In function 'mbuf_to_desc',
>>>      inlined from 'vhost_enqueue_async_packed'
>>> at ../lib/vhost/virtio_net.c:1826:6,
>>>      inlined from 'virtio_dev_rx_async_packed'
>>> at ../lib/vhost/virtio_net.c:1840:6,
>>>      inlined from 'virtio_dev_rx_async_submit_packed.constprop'
>>> at ../lib/vhost/virtio_net.c:1900:7:
>>> ../lib/vhost/virtio_net.c:1161:35: error: 'buf_vec[0].buf_len' may be 
>>> used
>>> uninitialized [-Werror=maybe-uninitialized]
>>>   1161 |         buf_len = buf_vec[vec_idx].buf_len;
>>>        |                   ~~~~~~~~~~~~~~~~^~~~~~~~
>>> ../lib/vhost/virtio_net.c: In function
>>> 'virtio_dev_rx_async_submit_packed.constprop':
>>> ../lib/vhost/virtio_net.c:1838:27: note: 'buf_vec' declared here
>>>   1838 |         struct buf_vector buf_vec[BUF_VECTOR_MAX];
>>>        |                           ^~~~~~~
>>> cc1: all warnings being treated as errors
>>>
>>> Actually, there are eight places to see the same codes in the file
>>> lib/vhost/virtio_net.c, and all these `buf_vec` arraies are
>>> initialized by sub-function calls under various conditions.
>>>
>>> Although It's hard to understand why gcc just emits warning at one
>>> of the eight places, adding validity checks for array length is
>>> reasonable and can also fix the warning.
>>>
>>> Signed-off-by: David Marchand <david.marchand@redhat.com>
>>> Signed-off-by: Min Zhou <zhoumin@loongson.cn>
>>> ---
>>>   lib/vhost/virtio_net.c | 5 ++++-
>>>   1 file changed, 4 insertions(+), 1 deletion(-)
>> Just want you to know that your patch is still pending because by 
>> accident
>> your fix is almost the same as a previous patch that fixes a real 
>> issue but
>> that patch is still in progress:
>>
>> http://patchwork.dpdk.org/project/dpdk/patch/20220802004938.23670-2-cfontana@suse.de/
>>
>> Thanks,
>> Chenbo
> 
> Thanks for your helpful reply.
> I think I can drop this patch if the patch you mentioned above could be 
> accepted.

I just sent another patch that fixes your build issue without having to
add runtime checks:

http://patches.dpdk.org/project/dpdk/patch/20221005144859.70717-1-maxime.coquelin@redhat.com/

Thanks,
Maxime
zhoumin Oct. 6, 2022, 3:56 a.m. UTC | #6 
Hi, Maxime,

Thanks a lot for your kind reply.

This patch is very useful and relieves the concerns in my heart.


On Wed, Oct 5, 2022 at 22:51, Maxime Coquelin wrote:

> Hi Min,
>
> On 9/26/22 05:25, zhoumin wrote:
>> Hi Chenbo,
>>
>>
>> On Mon, 26 Sep 2022, 10:57, Xia, Chenbo wrote:
>>> Hi Min,
>>>
>>>> -----Original Message-----
>>>> From: Min Zhou <zhoumin@loongson.cn>
>>>> Sent: Monday, August 29, 2022 4:29 PM
>>>> To: david.marchand@redhat.com; maxime.coquelin@redhat.com; Xia, Chenbo
>>>> <chenbo.xia@intel.com>; zhoumin@loongson.cn
>>>> Cc: dev@dpdk.org; maobibo@loongson.cn
>>>> Subject: [PATCH v2] vhost: fix build
>>>>
>>>> On CentOS 8 or Debian 10.4 systems using gcc 12.1 to cross
>>>> compile DPDK, gcc shows a following warning which will cause
>>>> build to fail when build is run with -werror:
>>>>
>>>> In function 'mbuf_to_desc',
>>>>      inlined from 'vhost_enqueue_async_packed'
>>>> at ../lib/vhost/virtio_net.c:1826:6,
>>>>      inlined from 'virtio_dev_rx_async_packed'
>>>> at ../lib/vhost/virtio_net.c:1840:6,
>>>>      inlined from 'virtio_dev_rx_async_submit_packed.constprop'
>>>> at ../lib/vhost/virtio_net.c:1900:7:
>>>> ../lib/vhost/virtio_net.c:1161:35: error: 'buf_vec[0].buf_len' may 
>>>> be used
>>>> uninitialized [-Werror=maybe-uninitialized]
>>>>   1161 |         buf_len = buf_vec[vec_idx].buf_len;
>>>>        |                   ~~~~~~~~~~~~~~~~^~~~~~~~
>>>> ../lib/vhost/virtio_net.c: In function
>>>> 'virtio_dev_rx_async_submit_packed.constprop':
>>>> ../lib/vhost/virtio_net.c:1838:27: note: 'buf_vec' declared here
>>>>   1838 |         struct buf_vector buf_vec[BUF_VECTOR_MAX];
>>>>        |                           ^~~~~~~
>>>> cc1: all warnings being treated as errors
>>>>
>>>> Actually, there are eight places to see the same codes in the file
>>>> lib/vhost/virtio_net.c, and all these `buf_vec` arraies are
>>>> initialized by sub-function calls under various conditions.
>>>>
>>>> Although It's hard to understand why gcc just emits warning at one
>>>> of the eight places, adding validity checks for array length is
>>>> reasonable and can also fix the warning.
>>>>
>>>> Signed-off-by: David Marchand <david.marchand@redhat.com>
>>>> Signed-off-by: Min Zhou <zhoumin@loongson.cn>
>>>> ---
>>>>   lib/vhost/virtio_net.c | 5 ++++-
>>>>   1 file changed, 4 insertions(+), 1 deletion(-)
>>> Just want you to know that your patch is still pending because by 
>>> accident
>>> your fix is almost the same as a previous patch that fixes a real 
>>> issue but
>>> that patch is still in progress:
>>>
>>> http://patchwork.dpdk.org/project/dpdk/patch/20220802004938.23670-2-cfontana@suse.de/ 
>>>
>>>
>>> Thanks,
>>> Chenbo
>>
>> Thanks for your helpful reply.
>> I think I can drop this patch if the patch you mentioned above could 
>> be accepted.
>
> I just sent another patch that fixes your build issue without having to
> add runtime checks:
>
> http://patches.dpdk.org/project/dpdk/patch/20221005144859.70717-1-maxime.coquelin@redhat.com/ 
>
>

OK, thanks. I have read the whole thread, which is a good job. I'd like 
thank you all for your hard work.


> Thanks,
> Maxime
>
zhoumin Oct. 6, 2022, 3:57 a.m. UTC | #7 
Hi, Maxime,

Thanks a lot for your kind reply.

This patch is very useful and relieves the concerns in my heart.


On Wed, Oct 5, 2022 at 22:51, Maxime Coquelin wrote:

> Hi Min,
>
> On 9/26/22 05:25, zhoumin wrote:
>> Hi Chenbo,
>>
>>
>> On Mon, 26 Sep 2022, 10:57, Xia, Chenbo wrote:
>>> Hi Min,
>>>
>>>> -----Original Message-----
>>>> From: Min Zhou <zhoumin@loongson.cn>
>>>> Sent: Monday, August 29, 2022 4:29 PM
>>>> To: david.marchand@redhat.com; maxime.coquelin@redhat.com; Xia, Chenbo
>>>> <chenbo.xia@intel.com>; zhoumin@loongson.cn
>>>> Cc: dev@dpdk.org; maobibo@loongson.cn
>>>> Subject: [PATCH v2] vhost: fix build
>>>>
>>>> On CentOS 8 or Debian 10.4 systems using gcc 12.1 to cross
>>>> compile DPDK, gcc shows a following warning which will cause
>>>> build to fail when build is run with -werror:
>>>>
>>>> In function 'mbuf_to_desc',
>>>>      inlined from 'vhost_enqueue_async_packed'
>>>> at ../lib/vhost/virtio_net.c:1826:6,
>>>>      inlined from 'virtio_dev_rx_async_packed'
>>>> at ../lib/vhost/virtio_net.c:1840:6,
>>>>      inlined from 'virtio_dev_rx_async_submit_packed.constprop'
>>>> at ../lib/vhost/virtio_net.c:1900:7:
>>>> ../lib/vhost/virtio_net.c:1161:35: error: 'buf_vec[0].buf_len' may 
>>>> be used
>>>> uninitialized [-Werror=maybe-uninitialized]
>>>>   1161 |         buf_len = buf_vec[vec_idx].buf_len;
>>>>        |                   ~~~~~~~~~~~~~~~~^~~~~~~~
>>>> ../lib/vhost/virtio_net.c: In function
>>>> 'virtio_dev_rx_async_submit_packed.constprop':
>>>> ../lib/vhost/virtio_net.c:1838:27: note: 'buf_vec' declared here
>>>>   1838 |         struct buf_vector buf_vec[BUF_VECTOR_MAX];
>>>>        |                           ^~~~~~~
>>>> cc1: all warnings being treated as errors
>>>>
>>>> Actually, there are eight places to see the same codes in the file
>>>> lib/vhost/virtio_net.c, and all these `buf_vec` arraies are
>>>> initialized by sub-function calls under various conditions.
>>>>
>>>> Although It's hard to understand why gcc just emits warning at one
>>>> of the eight places, adding validity checks for array length is
>>>> reasonable and can also fix the warning.
>>>>
>>>> Signed-off-by: David Marchand <david.marchand@redhat.com>
>>>> Signed-off-by: Min Zhou <zhoumin@loongson.cn>
>>>> ---
>>>>   lib/vhost/virtio_net.c | 5 ++++-
>>>>   1 file changed, 4 insertions(+), 1 deletion(-)
>>> Just want you to know that your patch is still pending because by 
>>> accident
>>> your fix is almost the same as a previous patch that fixes a real 
>>> issue but
>>> that patch is still in progress:
>>>
>>> http://patchwork.dpdk.org/project/dpdk/patch/20220802004938.23670-2-cfontana@suse.de/ 
>>>
>>>
>>> Thanks,
>>> Chenbo
>>
>> Thanks for your helpful reply.
>> I think I can drop this patch if the patch you mentioned above could 
>> be accepted.
>
> I just sent another patch that fixes your build issue without having to
> add runtime checks:
>
> http://patches.dpdk.org/project/dpdk/patch/20221005144859.70717-1-maxime.coquelin@redhat.com/ 
>
>

OK, thanks. I have read the whole thread, which is a good job. I'd like 
thank you all for your hard work.


> Thanks,
> Maxime
>
Thanks,

Min Zhou
diff mbox series

Patch

diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
index 35fa4670fd..99233f1759 100644
--- a/lib/vhost/virtio_net.c
+++ b/lib/vhost/virtio_net.c
@@ -1153,7 +1153,7 @@  mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq,
 	struct virtio_net_hdr_mrg_rxbuf tmp_hdr, *hdr = NULL;
 	struct vhost_async *async = vq->async;
 
-	if (unlikely(m == NULL))
+	if (unlikely(m == NULL || nr_vec == 0))
 		return -1;
 
 	buf_addr = buf_vec[vec_idx].buf_addr;
@@ -2673,6 +2673,9 @@  desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq,
 	struct vhost_async *async = vq->async;
 	struct async_inflight_info *pkts_info;
 
+	if (unlikely(nr_vec == 0))
+		return -1;
+
 	buf_addr = buf_vec[vec_idx].buf_addr;
 	buf_iova = buf_vec[vec_idx].buf_iova;
 	buf_len = buf_vec[vec_idx].buf_len;