VHDX Fixes for zero init

Minor cleanup.

Signed-off-by Jeff Cody <[email protected]>

---
  block/vhdx.c | 8 ++++----
  1 file changed, 4 insertions(+), 4 deletions(-)

block/vhdx.c +2 -2

index 12bfe75..f1e1e2e 100644
@@ -1109,8 +1109,8 @@ static coroutine_fn int vhdx_co_readv(BlockDriverState *bs, int64_t sector_num,
              /* check the payload block state */
              switch (s->bat[sinfo.bat_idx] & VHDX_BAT_STATE_BIT_MASK) {
              case PAYLOAD_BLOCK_NOT_PRESENT: /* fall through */
-            case PAYLOAD_BLOCK_UNDEFINED:   /* fall through */
-            case PAYLOAD_BLOCK_UNMAPPED:    /* fall through */
+            case PAYLOAD_BLOCK_UNDEFINED:
+            case PAYLOAD_BLOCK_UNMAPPED:
              case PAYLOAD_BLOCK_ZERO:
                  /* return zero */
                  qemu_iovec_memset(&hd_qiov, 0, 0, sinfo.bytes_avail);

block/vhdx.c +2 -2

-            case PAYLOAD_BLOCK_UNMAPPED:    /* fall through */
-            case PAYLOAD_BLOCK_UNDEFINED:   /* fall through */
+            case PAYLOAD_BLOCK_UNMAPPED:
+            case PAYLOAD_BLOCK_UNDEFINED:
                  bat_prior_offset = sinfo.file_offset;
                  ret = vhdx_allocate_block(bs, s, &sinfo.file_offset);
                  if (ret < 0) {

The 0.95 VHDX spec defined PAYLOAD_BLOCK_UNMAPPED to be 5. The 1.00 VHDX spec redefines PAYLOAD_BLOCK_UNMAPPED to be 3 instead.

The original value of 5 is now an undefined state in the spec, but it should be safe to treat it the same and return zeros for data read. This way, we can maintain compatibility with any images out in the wild that may have been created in accordance to the 0.95 spec.

Reported-by Kevin Wolf <[email protected]>

Signed-off-by Jeff Cody <[email protected]>

---
  block/vhdx.c | 3 ++-
  block/vhdx.h | 3 ++-
  2 files changed, 4 insertions(+), 2 deletions(-)

block/vhdx.c +1 -0

index f1e1e2e..bec10bd 100644
@@ -1111,6 +1111,7 @@ static coroutine_fn int vhdx_co_readv(BlockDriverState *bs, int64_t sector_num,
              case PAYLOAD_BLOCK_NOT_PRESENT: /* fall through */
              case PAYLOAD_BLOCK_UNDEFINED:
              case PAYLOAD_BLOCK_UNMAPPED:
+            case PAYLOAD_BLOCK_UNMAPPED_v095:
              case PAYLOAD_BLOCK_ZERO:
                  /* return zero */
                  qemu_iovec_memset(&hd_qiov, 0, 0, sinfo.bytes_avail);

block/vhdx.c +1 -0

                  /* fall through */
              case PAYLOAD_BLOCK_NOT_PRESENT: /* fall through */
              case PAYLOAD_BLOCK_UNMAPPED:
+            case PAYLOAD_BLOCK_UNMAPPED_v095:
              case PAYLOAD_BLOCK_UNDEFINED:
                  bat_prior_offset = sinfo.file_offset;
                  ret = vhdx_allocate_block(bs, s, &sinfo.file_offset);

block/vhdx.h +2 -1

index b4a12a0..7003ab7 100644
@@ -226,7 +226,8 @@ typedef struct QEMU_PACKED VHDXLogDataSector {
  #define PAYLOAD_BLOCK_NOT_PRESENT       0
  #define PAYLOAD_BLOCK_UNDEFINED         1
  #define PAYLOAD_BLOCK_ZERO              2
-#define PAYLOAD_BLOCK_UNMAPPED          5
+#define PAYLOAD_BLOCK_UNMAPPED          3
+#define PAYLOAD_BLOCK_UNMAPPED_v095     5
  #define PAYLOAD_BLOCK_FULLY_PRESENT     6
  #define PAYLOAD_BLOCK_PARTIALLY_PRESENT 7

The VHDX spec specifies that the default new block state is PAYLOAD_BLOCK_NOT_PRESENT for a dynamic VHDX image, and PAYLOAD_BLOCK_FULLY_PRESENT for a fixed VHDX image.

However, in order to create space-efficient VHDX images with qemu-img convert, it is desirable to be able to set has_zero_init to true for VHDX.

There is currently an option when creating VHDX images, to use block state ZERO for new blocks. However, this currently defaults to 'off'. In order to be able to eventually set has_zero_init to true for VHDX, this needs to default to 'on'.

This patch changes the default to 'on', and provides some help information to warn against setting it to 'off' when using qemu-img convert.

Signed-off-by Jeff Cody <[email protected]>

---
  block/vhdx.c  | 6 ++++--
  qemu-doc.texi | 6 +++++-
  2 files changed, 9 insertions(+), 3 deletions(-)

block/vhdx.c +1 -1

index bec10bd..ddefc2a 100644
@@ -1774,7 +1774,7 @@ static int vhdx_create(const char *filename, QemuOpts *opts, Error **errp)
      log_size = qemu_opt_get_size_del(opts, VHDX_BLOCK_OPT_LOG_SIZE, 0);
      block_size = qemu_opt_get_size_del(opts, VHDX_BLOCK_OPT_BLOCK_SIZE, 0);
      type = qemu_opt_get_del(opts, BLOCK_OPT_SUBFMT);
-    use_zero_blocks = qemu_opt_get_bool_del(opts, VHDX_BLOCK_OPT_ZERO, false);
+    use_zero_blocks = qemu_opt_get_bool_del(opts, VHDX_BLOCK_OPT_ZERO, true);
 
      if (image_size > VHDX_MAX_IMAGE_SIZE) {
          error_setg_errno(errp, EINVAL, "Image size too large; max of 64TB");

block/vhdx.c +0 -0

         },
         { NULL }
      }

qemu-doc.texi +0 -0

index ad418f8..aabe8df 100644

Now that new VHDX images will default to BAT block states of PAYLOAD_BLOCK_ZERO, we can indicate that VHDX has zero init.

Signed-off-by Jeff Cody <[email protected]>

---
  block/vhdx.c | 1 +
  1 file changed, 1 insertion(+)

Now that new VHDX images will default to BAT block states of PAYLOAD_BLOCK_ZERO, we can indicate that VHDX has zero init.

Signed-off-by Jeff Cody <[email protected]>

---
  block/vhdx.c | 1 +
  1 file changed, 1 insertion(+)

Hi,

I raised a bug https://bugs.launchpad.net/qemu/+bug/1399191 in the qemu bugzilla and immediately started getting these patch mails. But, I am not seeing any update in the bugzilla site. Till now I have got 4 patch mails regarding the issue. Please let me know which patch I should apply. Also let me know where can I get the patch files individually to be downloaded.

We are waiting for the patch to provide to our customers.

Thanks,
Amulya

Minor cleanup.

Signed-off-by Jeff Cody <[email protected]>

---
 block/vhdx.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

block/vhdx.c +2 -2

index 12bfe75..f1e1e2e 100644
@@ -1109,8 +1109,8 @@ static coroutine_fn int vhdx_co_readv(BlockDriverState *bs, int64_t sector_num,
             /* check the payload block state */
             switch (s->bat[sinfo.bat_idx] & VHDX_BAT_STATE_BIT_MASK) {
             case PAYLOAD_BLOCK_NOT_PRESENT: /* fall through */
-            case PAYLOAD_BLOCK_UNDEFINED:   /* fall through */
-            case PAYLOAD_BLOCK_UNMAPPED:    /* fall through */
+            case PAYLOAD_BLOCK_UNDEFINED:
+            case PAYLOAD_BLOCK_UNMAPPED:
             case PAYLOAD_BLOCK_ZERO:
                 /* return zero */
                 qemu_iovec_memset(&hd_qiov, 0, 0, sinfo.bytes_avail);

Hi,

I raised a bug https://bugs.launchpad.net/qemu/+bug/1399191 in the qemu bugzilla and immediately started getting these patch mails. But, I am not seeing any update in the bugzilla site.

Till now I have got 4 patch mails regarding the issue. Please let me know which patch I should apply. Also let me know where can I get the patch files individually to be downloaded.

We are waiting for the patch to provide to our customers.

Thanks,

Amulya

Hi Max,

We applied all the 5 patches from the mail chain I got since the last week. Please find attached the patches used by us.

We were unable to apply the patch3 as it failed with the following error

   # patch -p1 < patch3

patching file block/vhdx.c

patch: **** malformed patch at line 17:          error_setg_errno(errp, EINVAL, "Image size too large; max of 64TB"); @@ -1936,7 +1936,9 @@ static QemuOptsList vhdx_create_opts = {

Hence, we manually added the patch3 changes and recompiled the qemu. We then used the patched qemu-img to convert  our vmdk image to dynamic VHDX format. We found that the image created this time had a considerable decrease in its size from 50GB to 12GB.

However, when we deployed it into our SCVMM 2012, the import of the VHDX image failed with a "syntax error" as below

Information (10804)

Unable to import \\Test.com\Library\VHDs\Test-disk1.vhdx because of a syntax error in the file.

Please let us know if we missed anything.

Thanks,

Amulya

A couple of VHDX fixes in this series:
* updating the driver to reflect the 1.00 spec for the value of PAYLOAD_BLOCK_UNMAPPED (thanks Kevin)
* enabling VHDX to support zero init in qemu-img convert

1/4: compiling: e8718ae: block: vhdx - remove redundant comments
2/4: compiling: 8a4d2da: block: vhdx - update PAYLOAD_BLOCK_UNMAPPED value to match 1.00 spec 3/4: compiling: 44f08c2: block: vhdx - change .vhdx_create default block state to ZERO 4/4: compiling: 3eb0cd2: block: vhdx - set .bdrv_has_zero_init to bdrv_has_zero_init_1

Jeff Cody (4):
block: vhdx - remove redundant comments
block: vhdx - update PAYLOAD_BLOCK_UNMAPPED value to match 1.00 spec
block: vhdx - change .vhdx_create default block state to ZERO
block: vhdx - set .bdrv_has_zero_init to bdrv_has_zero_init_1

 block/vhdx.c  | 18 +++++++++++-------
 block/vhdx.h  |  3 ++-
 qemu-doc.texi |  6 +++++-
 3 files changed, 18 insertions(+), 9 deletions(-)

Signature · 1 line

1.9.3

Hi Max,

We applied all the 5 patches from the mail chain I got since the last week. Please find attached the patches used by us.

We were unable to apply the patch3 as it failed with the following
error

   # patch -p1 < patch3

patching file block/vhdx.c

patch: **** malformed patch at line 17:          error_setg_errno(errp, EINVAL, "Image size too large; max of 64TB"); @@ -1936,7 +1936,9 @@ static QemuOptsList vhdx_create_opts = {

Hence, we manually added the patch3 changes and recompiled the qemu. We then used the patched qemu-img to convert  our vmdk image to dynamic VHDX format. We found that the image created this time had a considerable decrease in its size from 50GB to 12GB.

However, when we deployed it into our SCVMM 2012, the import of the VHDX image failed with a "syntax error" as below

Information (10804)

Unable to import \\Test.com\Library\VHDs\Test-disk1.vhdx because of a syntax error in the file.

Please let us know if we missed anything.

Thanks,

Amulya

Quoted 7 lines · Jeff Cody

Minor cleanup.

Signed-off-by Jeff Cody <[email protected]>

---
 block/vhdx.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

block/vhdx.c static coroutine_fn int vhdx_co_readv(BlockDriverState *bs, int64_t sector_num, · line 1111

Quoted 8 lines

block/vhdx.c +0 -0

index 12bfe75..f1e1e2e 100644
@@ -1109,8 +1109,8 @@ static coroutine_fn int vhdx_co_readv(BlockDriverState *bs, int64_t sector_num,
             /* check the payload block state */
             switch (s->bat[sinfo.bat_idx] & VHDX_BAT_STATE_BIT_MASK) {
             case PAYLOAD_BLOCK_NOT_PRESENT: /* fall through */

block/vhdx.c static coroutine_fn int vhdx_co_writev(BlockDriverState *bs, int64_t sector_num, · line 1281

Quoted 10 lines

block/vhdx.c +2 -2

-            case PAYLOAD_BLOCK_UNDEFINED:   /* fall through */
-            case PAYLOAD_BLOCK_UNMAPPED:    /* fall through */
+            case PAYLOAD_BLOCK_UNDEFINED:
+            case PAYLOAD_BLOCK_UNMAPPED:
             case PAYLOAD_BLOCK_ZERO:
                 /* return zero */
                 qemu_iovec_memset(&hd_qiov, 0, 0, sinfo.bytes_avail);

@@ -1280,8 +1280,8 @@ static coroutine_fn int vhdx_co_writev(BlockDriverState *bs, int64_t sector_num,
 
                 /* fall through */

block/vhdx.c static coroutine_fn int vhdx_co_writev(BlockDriverState *bs, int64_t sector_num, · line 1282

Quoted 1 line

block/vhdx.c +0 -0

             case PAYLOAD_BLOCK_NOT_PRESENT: /* fall through */

block/vhdx.c static coroutine_fn int vhdx_co_writev(BlockDriverState *bs, int64_t sector_num, · line 1287

Quoted 7 lines

block/vhdx.c +2 -2

-            case PAYLOAD_BLOCK_UNMAPPED:    /* fall through */
-            case PAYLOAD_BLOCK_UNDEFINED:   /* fall through */
+            case PAYLOAD_BLOCK_UNMAPPED:
+            case PAYLOAD_BLOCK_UNDEFINED:
                 bat_prior_offset = sinfo.file_offset;
                 ret = vhdx_allocate_block(bs, s, &sinfo.file_offset);
                 if (ret < 0) {

Fall through comments are used by some static checkers. Not sure if all checkers are smart enough to propagate the comment from adjacent case statements. I would have left the comments alone but am okay with merging this.

Hi Max,

We applied all the 5 patches from the mail chain I got since the last week. Please find attached the patches used by us.

We were unable to apply the patch3 as it failed with the following
error

   # patch -p1 < patch3

patching file block/vhdx.c

patch: **** malformed patch at line 17:          error_setg_errno(errp, EINVAL, "Image size too large; max of 64TB"); @@ -1936,7 +1936,9 @@ static QemuOptsList vhdx_create_opts = {

Hence, we manually added the patch3 changes and recompiled the qemu. We then used the patched qemu-img to convert  our vmdk image to dynamic VHDX format. We found that the image created this time had a considerable decrease in its size from 50GB to 12GB.

However, when we deployed it into our SCVMM 2012, the import of the VHDX image failed with a "syntax error" as below

Information (10804)

Unable to import \\Test.com\Library\VHDs\Test-disk1.vhdx because of a syntax error in the file.

Please let us know if we missed anything.

Thanks,

Amulya

Hi Max, Jeff,

We were able to get the qemu patch files downloaded from the qemu patch site - https://patchwork.ozlabs.org and were able to apply the patches successfully without any errors. With the patches applied, we recompiled the qemu and converted the VDK vmdk to vhdx format and uploaded to the SCVMM Server. But it failed again with the syntax error as below:

Information (10804)
Unable to import \\TestServer\MSSCVMMLibrary\VHDs\Product-disk1.vhdx because of a syntax error in the file.

Please find my comments inline for your questions

Please let us know if there is anything else you need from us.

-----Original Message-----
From: Jeff Cody [mailto:[email protected]]
Sent: Friday, December 12, 2014 8:48 PM
To: Lokesha, Amulya
Cc: Max Reitz; [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH 4/4] block: vhdx - set .bdrv_has_zero_init to bdrv_has_zero_init_1

Quoted 3 lines · Amulya

Hi Max,

Please reply in-line, it makes it easier to follow technical discussions - thanks :)

Quoted 9 lines

We applied all the 5 patches from the mail chain I got since the last week. Please find attached the patches used by us.

We were unable to apply the patch3 as it failed with the following
error

Quoted 9 lines

   # patch -p1 < patch3

patching file block/vhdx.c

patch: **** malformed patch at line 17:          error_setg_errno(errp, EINVAL, "Image size too large; max of 64TB"); @@ -1936,7 +1936,9 @@ static QemuOptsList vhdx_create_opts = {

It looks like however you saved the patch file, it was corrupted. Looking at your attached patch 3, it split line 9 across 2 lines. Your patch also has whitespace differences from the patch I sent.

You also attached 5 patches - Why are you using patch 0? You should only be applying patches 1-4. This should not be causing any actual issues, however.
[Amulya]: First time we applied patches 1 to 4, created VHDX image and deployed to HyperV Server, but we got the same error. Then we took a fresh qemu source and applied patches 0 to 4 and deployed to HyperV and again got the same syntax error.

Are you using git for your qemu version? If so, 'git am' is the preferred method of applying the patches - just save each of the patch emails (the whole email should be fine), and run 'git am' on each file.

[Amulya] : No. We don't have a git repository for our team. Could you please let us know how to apply these patches without git. What is the difference in applying the patch directly and modifying the code directly? Does it have any impact?

Quoted 8 lines

Hence, we manually added the patch3 changes and recompiled the qemu. We then used the patched qemu-img to convert  our vmdk image to dynamic VHDX format. We found that the image created this time had a considerable decrease in its size from 50GB to 12GB.

Could you tell me the file size of the VMDK image you were converting? Is it roughly 12GB as well?
[Amulya] : No, the vmdk image which we used for conversion is just 1.4GB

Quoted 11 lines

However, when we deployed it into our SCVMM 2012, the import of the VHDX image failed with a "syntax error" as below

Information (10804)

Unable to import \\Test.com\Library\VHDs\Test-disk1.vhdx because of a syntax error in the file.

If you run qemu-img info on Test-disk1.vhdx, what does it say?

[Amulya] : The following is seen with qemu-img info

# qemu-img info Test-disk1.vhdx
image: Test-disk1.vhdx
file format: vhdx
virtual size: 50G (53687091200 bytes)
disk size: 3.4G
cluster_size: 16777216

The size of our images

# ls -ltrh
total 4.8G

(quoted diff) +0 -2

-rw-r--r-- 1 root root 1.4G Dec  9 10:48 Test-disk1.vmdk
-rw-r--r-- 1 root root  12G Dec 12 04:49 Test-disk1.vhdx

Quoted 4 lines

Please let us know if we missed anything.

Quoted 8 lines

Thanks,

Amulya

[...]

Jeff

Hi Max, Jeff,

We were able to get the qemu patch files downloaded from the qemu patch site - https://patchwork.ozlabs.org and were able to apply the patches successfully without any errors. With the patches applied, we recompiled the qemu and converted the VDK vmdk to vhdx format and uploaded to the SCVMM Server. But it failed again with the syntax error as below:

Information (10804)
Unable to import \\TestServer\MSSCVMMLibrary\VHDs\Product-disk1.vhdx because of a syntax error in the file.

Please find my comments inline for your questions

Please let us know if there is anything else you need from us.

-----Original Message-----
From: Jeff Cody [mailto:[email protected]]
Sent: Friday, December 12, 2014 8:48 PM
To: Lokesha, Amulya
Cc: Max Reitz; [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH 4/4] block: vhdx - set .bdrv_has_zero_init to bdrv_has_zero_init_1

Quoted 3 lines · Amulya

Hi Max,

Please reply in-line, it makes it easier to follow technical discussions - thanks :)

Quoted 9 lines

We applied all the 5 patches from the mail chain I got since the last week. Please find attached the patches used by us.

We were unable to apply the patch3 as it failed with the
following error

Quoted 9 lines

   # patch -p1 < patch3

patching file block/vhdx.c

patch: **** malformed patch at line 17:          error_setg_errno(errp, EINVAL, "Image size too large; max of 64TB"); @@ -1936,7 +1936,9 @@ static QemuOptsList vhdx_create_opts = {

It looks like however you saved the patch file, it was corrupted. Looking at your attached patch 3, it split line 9 across 2 lines. Your patch also has whitespace differences from the patch I sent.

You also attached 5 patches - Why are you using patch 0? You should only be applying patches 1-4. This should not be causing any actual issues, however.
[Amulya]: First time we applied patches 1 to 4, created VHDX image and deployed to HyperV Server, but we got the same error. Then we took a fresh qemu source and applied patches 0 to 4 and deployed to HyperV and again got the same syntax error.

Are you using git for your qemu version? If so, 'git am' is the preferred method of applying the patches - just save each of the patch emails (the whole email should be fine), and run 'git am' on each file.

[Amulya] : No. We don't have a git repository for our team. Could you please let us know how to apply these patches without git. What is the difference in applying the patch directly and modifying the code directly? Does it have any impact?

Quoted 8 lines

Hence, we manually added the patch3 changes and recompiled the qemu. We then used the patched qemu-img to convert  our vmdk image to dynamic VHDX format. We found that the image created this time had a considerable decrease in its size from 50GB to 12GB.

Could you tell me the file size of the VMDK image you were converting? Is it roughly 12GB as well?
[Amulya] : No, the vmdk image which we used for conversion is just 1.4GB

Quoted 11 lines

However, when we deployed it into our SCVMM 2012, the import of the VHDX image failed with a "syntax error" as below

Information (10804)

Unable to import \\Test.com\Library\VHDs\Test-disk1.vhdx because of a syntax error in the file.

If you run qemu-img info on Test-disk1.vhdx, what does it say?

[Amulya] : The following is seen with qemu-img info

# qemu-img info Test-disk1.vhdx
image: Test-disk1.vhdx
file format: vhdx
virtual size: 50G (53687091200 bytes)
disk size: 3.4G
cluster_size: 16777216

The size of our images

# ls -ltrh
total 4.8G

(quoted diff) +0 -2

-rw-r--r-- 1 root root 1.4G Dec  9 10:48 Test-disk1.vmdk
-rw-r--r-- 1 root root  12G Dec 12 04:49 Test-disk1.vhdx

Quoted 4 lines

Please let us know if we missed anything.

Quoted 8 lines

Thanks,

Amulya

[...]

Jeff

-----Original Message-----
From: Jeff Cody [mailto:[email protected]]
Sent: Wednesday, December 17, 2014 5:44 PM
To: Lokesha, Amulya
Cc: Max Reitz; [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH 4/4] block: vhdx - set .bdrv_has_zero_init to bdrv_has_zero_init_1

Quoted 12 lines · Amulya

Hi Max, Jeff,

We were able to get the qemu patch files downloaded from the qemu patch site - https://patchwork.ozlabs.org and were able to apply the patches successfully without any errors. With the patches applied, we recompiled the qemu and converted the VDK vmdk to vhdx format and uploaded to the SCVMM Server. But it failed again with the syntax error as below:

Information (10804)
Unable to import \\TestServer\MSSCVMMLibrary\VHDs\Product-disk1.vhdx because of a syntax error in the file.

Please find my comments inline for your questions

Please let us know if there is anything else you need from us.

Amulya,

I will try to test this on Windows Server, and see if I can reproduce what you are seeing.

-Jeff

Hi Jeff,

Any updates on this? Were you able to test it

Thanks,
Amulya

Quoted 109 lines

-----Original Message-----
From: Jeff Cody [mailto:[email protected]]
Sent: Friday, December 12, 2014 8:48 PM
To: Lokesha, Amulya
Cc: Max Reitz; [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH 4/4] block: vhdx - set .bdrv_has_zero_init to bdrv_has_zero_init_1

Quoted 3 lines · Amulya

Hi Max,

Please reply in-line, it makes it easier to follow technical discussions - thanks :)

Quoted 9 lines

We applied all the 5 patches from the mail chain I got since the last week. Please find attached the patches used by us.

We were unable to apply the patch3 as it failed with the
following error

Quoted 9 lines

   # patch -p1 < patch3

patching file block/vhdx.c

patch: **** malformed patch at line 17:          error_setg_errno(errp, EINVAL, "Image size too large; max of 64TB"); @@ -1936,7 +1936,9 @@ static QemuOptsList vhdx_create_opts = {

It looks like however you saved the patch file, it was corrupted. Looking at your attached patch 3, it split line 9 across 2 lines. Your patch also has whitespace differences from the patch I sent.

You also attached 5 patches - Why are you using patch 0? You should only be applying patches 1-4. This should not be causing any actual issues, however.
[Amulya]: First time we applied patches 1 to 4, created VHDX image and deployed to HyperV Server, but we got the same error. Then we took a fresh qemu source and applied patches 0 to 4 and deployed to HyperV and again got the same syntax error.

Are you using git for your qemu version? If so, 'git am' is the preferred method of applying the patches - just save each of the patch emails (the whole email should be fine), and run 'git am' on each file.

[Amulya] : No. We don't have a git repository for our team. Could you please let us know how to apply these patches without git. What is the difference in applying the patch directly and modifying the code directly? Does it have any impact?

Quoted 8 lines

Hence, we manually added the patch3 changes and recompiled the qemu. We then used the patched qemu-img to convert  our vmdk image to dynamic VHDX format. We found that the image created this time had a considerable decrease in its size from 50GB to 12GB.

Could you tell me the file size of the VMDK image you were converting? Is it roughly 12GB as well?
[Amulya] : No, the vmdk image which we used for conversion is just 1.4GB

Quoted 11 lines

However, when we deployed it into our SCVMM 2012, the import of the VHDX image failed with a "syntax error" as below

Information (10804)

Unable to import \\Test.com\Library\VHDs\Test-disk1.vhdx because of a syntax error in the file.

If you run qemu-img info on Test-disk1.vhdx, what does it say?

[Amulya] : The following is seen with qemu-img info

# qemu-img info Test-disk1.vhdx
image: Test-disk1.vhdx
file format: vhdx
virtual size: 50G (53687091200 bytes)
disk size: 3.4G
cluster_size: 16777216

The size of our images

# ls -ltrh
total 4.8G

(quoted diff) +0 -2

-rw-r--r-- 1 root root 1.4G Dec  9 10:48 Test-disk1.vmdk
-rw-r--r-- 1 root root  12G Dec 12 04:49 Test-disk1.vhdx

Quoted 4 lines

Please let us know if we missed anything.

Quoted 8 lines

Thanks,

Amulya

[...]

Jeff

-----Original Message-----
From: Jeff Cody [mailto:[email protected]]
Sent: Wednesday, December 17, 2014 5:44 PM
To: Lokesha, Amulya
Cc: Max Reitz; [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH 4/4] block: vhdx - set .bdrv_has_zero_init to bdrv_has_zero_init_1

Quoted 12 lines · Amulya

Hi Max, Jeff,

We were able to get the qemu patch files downloaded from the qemu patch site - https://patchwork.ozlabs.org and were able to apply the patches successfully without any errors. With the patches applied, we recompiled the qemu and converted the VDK vmdk to vhdx format and uploaded to the SCVMM Server. But it failed again with the syntax error as below:

Information (10804)
Unable to import \\TestServer\MSSCVMMLibrary\VHDs\Product-disk1.vhdx because of a syntax error in the file.

Please find my comments inline for your questions

Please let us know if there is anything else you need from us.

Amulya,

I will try to test this on Windows Server, and see if I can reproduce what you are seeing.

-Jeff

Hi Jeff,

Any updates on this? Were you able to test it

Thanks,
Amulya

Quoted 107 lines

-----Original Message-----
From: Jeff Cody [mailto:[email protected]]
Sent: Friday, December 12, 2014 8:48 PM
To: Lokesha, Amulya
Cc: Max Reitz; [email protected]; [email protected]; [email protected]
Subject: Re: [PATCH 4/4] block: vhdx - set .bdrv_has_zero_init to bdrv_has_zero_init_1

Quoted 3 lines · Amulya

Hi Max,

Please reply in-line, it makes it easier to follow technical discussions - thanks :)

Quoted 9 lines

We applied all the 5 patches from the mail chain I got since the last week. Please find attached the patches used by us.

We were unable to apply the patch3 as it failed with the
following error

Quoted 9 lines

   # patch -p1 < patch3

patching file block/vhdx.c

patch: **** malformed patch at line 17:          error_setg_errno(errp, EINVAL, "Image size too large; max of 64TB"); @@ -1936,7 +1936,9 @@ static QemuOptsList vhdx_create_opts = {

It looks like however you saved the patch file, it was corrupted. Looking at your attached patch 3, it split line 9 across 2 lines. Your patch also has whitespace differences from the patch I sent.

You also attached 5 patches - Why are you using patch 0? You should only be applying patches 1-4. This should not be causing any actual issues, however.
[Amulya]: First time we applied patches 1 to 4, created VHDX image and deployed to HyperV Server, but we got the same error. Then we took a fresh qemu source and applied patches 0 to 4 and deployed to HyperV and again got the same syntax error.

Are you using git for your qemu version? If so, 'git am' is the preferred method of applying the patches - just save each of the patch emails (the whole email should be fine), and run 'git am' on each file.

[Amulya] : No. We don't have a git repository for our team. Could you please let us know how to apply these patches without git. What is the difference in applying the patch directly and modifying the code directly? Does it have any impact?

Quoted 8 lines

Hence, we manually added the patch3 changes and recompiled the qemu. We then used the patched qemu-img to convert  our vmdk image to dynamic VHDX format. We found that the image created this time had a considerable decrease in its size from 50GB to 12GB.

Could you tell me the file size of the VMDK image you were converting? Is it roughly 12GB as well?
[Amulya] : No, the vmdk image which we used for conversion is just 1.4GB

Quoted 11 lines

However, when we deployed it into our SCVMM 2012, the import of the VHDX image failed with a "syntax error" as below

Information (10804)

Unable to import \\Test.com\Library\VHDs\Test-disk1.vhdx because of a syntax error in the file.

If you run qemu-img info on Test-disk1.vhdx, what does it say?

[Amulya] : The following is seen with qemu-img info

# qemu-img info Test-disk1.vhdx
image: Test-disk1.vhdx
file format: vhdx
virtual size: 50G (53687091200 bytes) disk size: 3.4G
cluster_size: 16777216

The size of our images

# ls -ltrh
total 4.8G

(quoted diff) +0 -2

-rw-r--r-- 1 root root 1.4G Dec  9 10:48 Test-disk1.vmdk
-rw-r--r-- 1 root root  12G Dec 12 04:49 Test-disk1.vhdx

Quoted 4 lines

Please let us know if we missed anything.

Quoted 8 lines

Thanks,

Amulya

[...]

Jeff