[Qemu-devel] [PATCH v2 0/1] qemu: fix the bug reported by qemu-iotests case 055

[Qemu-devel] [PATCH v2 0/1] qemu: fix the bug reported by qemu-iotests case 055

[QingFeng Hao]

Hi all,
v2 includes changes due to review comments by Kevin Wolf(thanks to Kevin).

v2:
* Add endian conversion for lba field in vmdk_write_extent.
  Based on master's commit:
  00227fefd205: Update version for v2.8.0-rc1 release

v1:
* Add patch to fix the bug reported by qemu-iotests case 055.
  Jing Liu and I found the cause was in vmdk and the fix is in the
  followed patch.
  Based on master's commit:
  00227fefd205: Update version for v2.8.0-rc1 release

  Upstream master's qemu-iotests case 055 reports the following error:
  +qemu-img: Error while reading offset 0 of tests/qemu-iotests/scratch/blockdev-target.img: Invalid argument
  +qemu-img: Error while reading offset 0 of tests/qemu-iotests/scratch/blockdev-target.img: Invalid argument
  +qemu-img: Error while reading offset 0 of tests/qemu-iotests/scratch/blockdev-target.img: Invalid argument
  +qemu-img: Error while reading offset 0 of tests/qemu-iotests/scratch/blockdev-target.img: Invalid argument 

Thanks!

QingFeng Hao (1):
  block/vmdk: Fix the endian problem of buf_len and lba

 block/vmdk.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

-- 
2.8.4

[Qemu-devel] [PATCH v2 1/1] block/vmdk: Fix the endian problem of buf_len and lba

[QingFeng Hao]

The problem was triggered by qemu-iotests case 055. It failed when it
was comparing the compressed vmdk image with original test.img.

The cause is that buf_len in vmdk_write_extent wasn't converted to
little-endian before it was stored to disk. But later vmdk_read_extent
read it and converted it from little-endian to cpu endian.
If the cpu is big-endian like s390, the problem will happen and
the data length read by vmdk_read_extent will become invalid!
The fix is to add the conversion in vmdk_write_extent, meanwhile,
repair the endianness problem of lba field which shall also be converted
to little-endian before storing to disk.

Cc: qemu-stable@nongnu.org
Signed-off-by: QingFeng Hao <haoqf@linux.vnet.ibm.com>
Signed-off-by: Jing Liu <liujbjl@linux.vnet.ibm.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
 block/vmdk.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/block/vmdk.c b/block/vmdk.c
index a11c27a..26e5f95 100644
--- a/block/vmdk.c
+++ b/block/vmdk.c
@@ -1354,8 +1354,8 @@ static int vmdk_write_extent(VmdkExtent *extent, int64_t cluster_offset,
             goto out;
         }
 
-        data->lba = offset >> BDRV_SECTOR_BITS;
-        data->size = buf_len;
+        data->lba = cpu_to_le64(offset >> BDRV_SECTOR_BITS);
+        data->size = cpu_to_le32(buf_len);
 
         n_bytes = buf_len + sizeof(VmdkGrainMarker);
         iov = (struct iovec) {
-- 
2.8.4

Re: [Qemu-devel] [Qemu-stable] [PATCH v2 1/1] block/vmdk: Fix the endian problem of buf_len and lba

[Fam Zheng]

On Sat, 11/26 06:46, QingFeng Hao wrote:
> The problem was triggered by qemu-iotests case 055. It failed when it
> was comparing the compressed vmdk image with original test.img.
> 
> The cause is that buf_len in vmdk_write_extent wasn't converted to
> little-endian before it was stored to disk. But later vmdk_read_extent
> read it and converted it from little-endian to cpu endian.
> If the cpu is big-endian like s390, the problem will happen and
> the data length read by vmdk_read_extent will become invalid!
> The fix is to add the conversion in vmdk_write_extent, meanwhile,
> repair the endianness problem of lba field which shall also be converted
> to little-endian before storing to disk.
> 
> Cc: qemu-stable@nongnu.org
> Signed-off-by: QingFeng Hao <haoqf@linux.vnet.ibm.com>
> Signed-off-by: Jing Liu <liujbjl@linux.vnet.ibm.com>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
>  block/vmdk.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/block/vmdk.c b/block/vmdk.c
> index a11c27a..26e5f95 100644
> --- a/block/vmdk.c
> +++ b/block/vmdk.c
> @@ -1354,8 +1354,8 @@ static int vmdk_write_extent(VmdkExtent *extent, int64_t cluster_offset,
>              goto out;
>          }
>  
> -        data->lba = offset >> BDRV_SECTOR_BITS;
> -        data->size = buf_len;
> +        data->lba = cpu_to_le64(offset >> BDRV_SECTOR_BITS);
> +        data->size = cpu_to_le32(buf_len);
>  
>          n_bytes = buf_len + sizeof(VmdkGrainMarker);
>          iov = (struct iovec) {
> -- 
> 2.8.4
> 
> 

Reviewed-by: Fam Zheng <famz@redhat.com>

Re: [Qemu-devel] [Qemu-stable] [PATCH v2 1/1] block/vmdk: Fix the endian problem of buf_len and lba

[Hao QingFeng]

在 2016-11-26 19:46, Fam Zheng 写道:
> On Sat, 11/26 06:46, QingFeng Hao wrote:
>> The problem was triggered by qemu-iotests case 055. It failed when it
>> was comparing the compressed vmdk image with original test.img.
>>
>> The cause is that buf_len in vmdk_write_extent wasn't converted to
>> little-endian before it was stored to disk. But later vmdk_read_extent
>> read it and converted it from little-endian to cpu endian.
>> If the cpu is big-endian like s390, the problem will happen and
>> the data length read by vmdk_read_extent will become invalid!
>> The fix is to add the conversion in vmdk_write_extent, meanwhile,
>> repair the endianness problem of lba field which shall also be converted
>> to little-endian before storing to disk.
>>
>> Cc: qemu-stable@nongnu.org
>> Signed-off-by: QingFeng Hao <haoqf@linux.vnet.ibm.com>
>> Signed-off-by: Jing Liu <liujbjl@linux.vnet.ibm.com>
>> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
>> ---
>>   block/vmdk.c | 4 ++--
>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/block/vmdk.c b/block/vmdk.c
>> index a11c27a..26e5f95 100644
>> --- a/block/vmdk.c
>> +++ b/block/vmdk.c
>> @@ -1354,8 +1354,8 @@ static int vmdk_write_extent(VmdkExtent *extent, int64_t cluster_offset,
>>               goto out;
>>           }
>>   
>> -        data->lba = offset >> BDRV_SECTOR_BITS;
>> -        data->size = buf_len;
>> +        data->lba = cpu_to_le64(offset >> BDRV_SECTOR_BITS);
>> +        data->size = cpu_to_le32(buf_len);
>>   
>>           n_bytes = buf_len + sizeof(VmdkGrainMarker);
>>           iov = (struct iovec) {
>> -- 
>> 2.8.4
>>
>>
> Reviewed-by: Fam Zheng <famz@redhat.com>
Thanks!

-- 
QingFeng Hao(Robin)

Re: [Qemu-devel] [PATCH v2 1/1] block/vmdk: Fix the endian problem of buf_len and lba

[liujing]

Hi QingFeng,


I just have a question that whether the marker->data
need convert?

I've no idea, just suddenly realized this question.

Jing

On 11/26/2016 01:46 PM, QingFeng Hao wrote:
> The problem was triggered by qemu-iotests case 055. It failed when it
> was comparing the compressed vmdk image with original test.img.
>
> The cause is that buf_len in vmdk_write_extent wasn't converted to
> little-endian before it was stored to disk. But later vmdk_read_extent
> read it and converted it from little-endian to cpu endian.
> If the cpu is big-endian like s390, the problem will happen and
> the data length read by vmdk_read_extent will become invalid!
> The fix is to add the conversion in vmdk_write_extent, meanwhile,
> repair the endianness problem of lba field which shall also be converted
> to little-endian before storing to disk.
>
> Cc: qemu-stable@nongnu.org
> Signed-off-by: QingFeng Hao <haoqf@linux.vnet.ibm.com>
> Signed-off-by: Jing Liu <liujbjl@linux.vnet.ibm.com>
> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
> ---
>   block/vmdk.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/block/vmdk.c b/block/vmdk.c
> index a11c27a..26e5f95 100644
> --- a/block/vmdk.c
> +++ b/block/vmdk.c
> @@ -1354,8 +1354,8 @@ static int vmdk_write_extent(VmdkExtent *extent, int64_t cluster_offset,
>               goto out;
>           }
>
> -        data->lba = offset >> BDRV_SECTOR_BITS;
> -        data->size = buf_len;
> +        data->lba = cpu_to_le64(offset >> BDRV_SECTOR_BITS);
> +        data->size = cpu_to_le32(buf_len);
>
>           n_bytes = buf_len + sizeof(VmdkGrainMarker);
>           iov = (struct iovec) {

Re: [Qemu-devel] [PATCH v2 1/1] block/vmdk: Fix the endian problem of buf_len and lba

[Hao QingFeng]

在 2016-11-28 15:56, liujing 写道:
> Hi QingFeng,
>
>
> I just have a question that whether the marker->data
> need convert?
>
> I've no idea, just suddenly realized this question.
>
nope, the data is type of char * for the compressed data stream, so no 
endian issue.
thanks.
> Jing
>
> On 11/26/2016 01:46 PM, QingFeng Hao wrote:
>> The problem was triggered by qemu-iotests case 055. It failed when it
>> was comparing the compressed vmdk image with original test.img.
>>
>> The cause is that buf_len in vmdk_write_extent wasn't converted to
>> little-endian before it was stored to disk. But later vmdk_read_extent
>> read it and converted it from little-endian to cpu endian.
>> If the cpu is big-endian like s390, the problem will happen and
>> the data length read by vmdk_read_extent will become invalid!
>> The fix is to add the conversion in vmdk_write_extent, meanwhile,
>> repair the endianness problem of lba field which shall also be converted
>> to little-endian before storing to disk.
>>
>> Cc: qemu-stable@nongnu.org
>> Signed-off-by: QingFeng Hao <haoqf@linux.vnet.ibm.com>
>> Signed-off-by: Jing Liu <liujbjl@linux.vnet.ibm.com>
>> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
>> ---
>>   block/vmdk.c | 4 ++--
>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/block/vmdk.c b/block/vmdk.c
>> index a11c27a..26e5f95 100644
>> --- a/block/vmdk.c
>> +++ b/block/vmdk.c
>> @@ -1354,8 +1354,8 @@ static int vmdk_write_extent(VmdkExtent 
>> *extent, int64_t cluster_offset,
>>               goto out;
>>           }
>>
>> -        data->lba = offset >> BDRV_SECTOR_BITS;
>> -        data->size = buf_len;
>> +        data->lba = cpu_to_le64(offset >> BDRV_SECTOR_BITS);
>> +        data->size = cpu_to_le32(buf_len);
>>
>>           n_bytes = buf_len + sizeof(VmdkGrainMarker);
>>           iov = (struct iovec) {
>

-- 
QingFeng Hao(Robin)

Re: [Qemu-devel] [PATCH v2 0/1] qemu: fix the bug reported by qemu-iotests case 055

[Hao QingFeng]

Sorry, missed Cc qemu-stable in this patch.


在 2016-11-26 13:46, QingFeng Hao 写道:
> Hi all,
> v2 includes changes due to review comments by Kevin Wolf(thanks to Kevin).
>
> v2:
> * Add endian conversion for lba field in vmdk_write_extent.
>    Based on master's commit:
>    00227fefd205: Update version for v2.8.0-rc1 release
>
> v1:
> * Add patch to fix the bug reported by qemu-iotests case 055.
>    Jing Liu and I found the cause was in vmdk and the fix is in the
>    followed patch.
>    Based on master's commit:
>    00227fefd205: Update version for v2.8.0-rc1 release
>
>    Upstream master's qemu-iotests case 055 reports the following error:
>    +qemu-img: Error while reading offset 0 of tests/qemu-iotests/scratch/blockdev-target.img: Invalid argument
>    +qemu-img: Error while reading offset 0 of tests/qemu-iotests/scratch/blockdev-target.img: Invalid argument
>    +qemu-img: Error while reading offset 0 of tests/qemu-iotests/scratch/blockdev-target.img: Invalid argument
>    +qemu-img: Error while reading offset 0 of tests/qemu-iotests/scratch/blockdev-target.img: Invalid argument
>
> Thanks!
>
> QingFeng Hao (1):
>    block/vmdk: Fix the endian problem of buf_len and lba
>
>   block/vmdk.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
>

-- 
QingFeng Hao(Robin)