Skip to contentSkip to conversation

block: deprecate the sheepdog driver

PATCH Superseded Daniel P. Berrangé 8 participants 15 messages Sep 22, 2020 – Sep 23, 2020 Block layer

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover block: deprecate the sheepdog driver

2 years back I proposed dropping the sheepdog mailing list from the MAINTAINERS file, but somehow the patch never got picked up:

https://lists.gnu.org/archive/html/qemu-block/2018-03/msg01048.html

So here I am with the same patch again.

This time I go further and deprecate the sheepdog driver entirely. See the rationale in the second patch commit message.

Daniel P. Berrangé (2):
block: drop moderated sheepdog mailing list from MAINTAINERS file
block: deprecate the sheepdog block driver

 MAINTAINERS                |  1 -
 block/sheepdog.c           | 15 +++++++++++++++
 configure                  |  5 +++--
 docs/system/deprecated.rst |  9 +++++++++
 4 files changed, 27 insertions(+), 3 deletions(-)
Signature · 1 line
2.26.2
#1 lore raw

PATCH v2 1/2 block: drop moderated sheepdog mailing list from MAINTAINERS file

The sheepdog mailing list is setup to stop and queue messages from non-subscribers, pending moderator approval. Unfortunately it seems that the moderation queue is not actively dealt with. Even when messages are approved, the sender is never added to the whitelist, so every future mail from the same sender continues to get stopped for moderation.

MAINTAINERS entries should be responsive and not unneccessarily block mails from QEMU contributors, so drop the sheepdog mailing list.

---
 MAINTAINERS | 1 -
 1 file changed, 1 deletion(-)
MAINTAINERS +0 -1
index 3d17cad19a..8e8a4fb0a8 100644
@@ -2852,7 +2852,6 @@ F: block/rbd.c
 Sheepdog
 M: Liu Yuan <[email protected]>
 L: [email protected]
-L: [email protected]
 S: Odd Fixes
 F: block/sheepdog.c
 
Signature · 1 line
2.26.2
Reviewed-by on v2 1/2
#2 lore raw

Re: [PATCH v2 1/2] block: drop moderated sheepdog mailing list from MAINTAINERS file

Quoted 20 lines · Daniel P. Berrangé

The sheepdog mailing list is setup to stop and queue messages from non-subscribers, pending moderator approval. Unfortunately it seems that the moderation queue is not actively dealt with. Even when messages are approved, the sender is never added to the whitelist, so every future mail from the same sender continues to get stopped for moderation.

MAINTAINERS entries should be responsive and not unneccessarily block mails from QEMU contributors, so drop the sheepdog mailing list.

---
 MAINTAINERS | 1 -
 1 file changed, 1 deletion(-)
MAINTAINERS +0 -0
index 3d17cad19a..8e8a4fb0a8 100644

MAINTAINERS F: block/rbd.c · line 2856

Quoted 6 lines
MAINTAINERS +0 -1
@@ -2852,7 +2852,6 @@ F: block/rbd.c
 Sheepdog
 M: Liu Yuan <[email protected]>
 L: [email protected]
-L: [email protected]
 S: Odd Fixes
 F: block/sheepdog.c
Reviewed-by on v2 1/2
#3 lore raw

Re: [PATCH v2 1/2] block: drop moderated sheepdog mailing list from MAINTAINERS file

In reply to Daniel P. Berrangé
Quoted 11 lines · Daniel P. Berrangé

The sheepdog mailing list is setup to stop and queue messages from non-subscribers, pending moderator approval. Unfortunately it seems that the moderation queue is not actively dealt with. Even when messages are approved, the sender is never added to the whitelist, so every future mail from the same sender continues to get stopped for moderation.

MAINTAINERS entries should be responsive and not unneccessarily block mails from QEMU contributors, so drop the sheepdog mailing list.

Signature · 2 lines
Best regards,
Vladimir
#4 lore raw

PATCH v2 2/2 block: deprecate the sheepdog block driver

In reply to Daniel P. Berrangé

This thread from a little over a year ago:

http://lists.wpkg.org/pipermail/sheepdog/2019-March/thread.html

states that sheepdog is no longer actively developed. The only mentioned users are some companies who are said to have it for legacy reasons with plans to replace it by Ceph. There is talk about cutting out existing features to turn it into a simple demo of how to write a distributed block service. There is no evidence of anyone working on that idea:

https://github.com/sheepdog/sheepdog/commits/master

No real commits to git since Jan 2018, and before then just some minor technical debt cleanup..

There is essentially no activity on the mailing list aside from patches to QEMU that get CC'd due to our MAINTAINERS entry.

Fedora packages for sheepdog failed to build from upstream source because of the more strict linker that no longer merges duplicate global symbols. Fedora patches it to add the missing "extern" annotations and presumably other distros do to, but upstream source remains broken.

There is only basic compile testing, no functional testing of the driver.

Since there are no build pre-requisites the sheepdog driver is currently enabled unconditionally. This would result in configure issuing a deprecation warning by default for all users. Thus the configure default is changed to disable it, requiring users to pass --enable-sheepdog to build the driver.

---
 block/sheepdog.c           | 15 +++++++++++++++
 configure                  |  5 +++--
 docs/system/deprecated.rst |  9 +++++++++
 3 files changed, 27 insertions(+), 2 deletions(-)
block/sheepdog.c +15 -0
index cbbebc1aaf..7f68bd6a1a 100644
@@ -242,6 +242,17 @@ typedef struct SheepdogInode {
  */
 #define FNV1A_64_INIT ((uint64_t)0xcbf29ce484222325ULL)
 
+static void deprecation_warning(void)
+{
+    static bool warned = false;
+
+    if (!warned) {
+        warn_report("the sheepdog block driver is deprecated and will be "
+                    "removed in a future release");
+        warned = true;
+    }
+}
+
 /*
  * 64 bit Fowler/Noll/Vo FNV-1a hash code
  */
@@ -1548,6 +1559,8 @@ static int sd_open(BlockDriverState *bs, QDict *options, int flags,
     char *buf = NULL;
     QemuOpts *opts;
 
+    deprecation_warning();
+
     s->bs = bs;
     s->aio_context = bdrv_get_aio_context(bs);
 
@@ -2007,6 +2020,8 @@ static int sd_co_create(BlockdevCreateOptions *options, Error **errp)
 
     assert(options->driver == BLOCKDEV_DRIVER_SHEEPDOG);
 
+    deprecation_warning();
+
     s = g_new0(BDRVSheepdogState, 1);
 
     /* Steal SocketAddress from QAPI, set NULL to prevent double free */
configure +3 -2
index 7564479008..c6af83f2e6 100755
@@ -533,7 +533,7 @@ vdi="yes"
 vvfat="yes"
 qed="yes"
 parallels="yes"
-sheepdog="yes"
+sheepdog="no"
 libxml2=""
 debug_mutex="no"
 libpmem=""
@@ -1941,7 +1941,7 @@ disabled with --disable-FEATURE, default is enabled if available:
   vvfat           vvfat image format support
   qed             qed image format support
   parallels       parallels image format support
-  sheepdog        sheepdog block driver support
+  sheepdog        sheepdog block driver support (deprecated)
   crypto-afalg    Linux AF_ALG crypto backend driver
   capstone        capstone disassembler support
   debug-mutex     mutex debugging support
@@ -7350,6 +7350,7 @@ if test "$parallels" = "yes" ; then
   echo "CONFIG_PARALLELS=y" >> $config_host_mak
 fi
 if test "$sheepdog" = "yes" ; then
+  add_to deprecated_features "sheepdog"
   echo "CONFIG_SHEEPDOG=y" >> $config_host_mak
 fi
 if test "$pty_h" = "yes" ; then
docs/system/deprecated.rst +9 -0
index 0cb8b01424..49b9f4b02e 100644
@@ -405,6 +405,15 @@ The above, converted to the current supported format::
 
   json:{"file.driver":"rbd", "file.pool":"rbd", "file.image":"name"}
 
+``sheepdog`` driver (since 5.2.0)
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The ``sheepdog`` block device driver is deprecated. The corresponding upstream
+server project is no longer actively maintained. Users are recommended to switch
+to an alternative distributed block device driver such as RBD. The
+``qemu-img convert`` command can be used to liberate existing data by moving
+it out of sheepdog volumes into an alternative storage backend.
+
 linux-user mode CPUs
 --------------------
 
Signature · 1 line
2.26.2
Reviewed-by on v2 2/2
#5 lore raw

Re: [PATCH v2 2/2] block: deprecate the sheepdog block driver

Quoted 33 lines · Daniel P. Berrangé

This thread from a little over a year ago:

http://lists.wpkg.org/pipermail/sheepdog/2019-March/thread.html

states that sheepdog is no longer actively developed. The only mentioned users are some companies who are said to have it for legacy reasons with plans to replace it by Ceph. There is talk about cutting out existing features to turn it into a simple demo of how to write a distributed block service. There is no evidence of anyone working on that idea:

https://github.com/sheepdog/sheepdog/commits/master

No real commits to git since Jan 2018, and before then just some minor technical debt cleanup..

There is essentially no activity on the mailing list aside from patches to QEMU that get CC'd due to our MAINTAINERS entry.

Fedora packages for sheepdog failed to build from upstream source because of the more strict linker that no longer merges duplicate global symbols. Fedora patches it to add the missing "extern" annotations and presumably other distros do to, but upstream source remains broken.

There is only basic compile testing, no functional testing of the driver.

Since there are no build pre-requisites the sheepdog driver is currently enabled unconditionally. This would result in configure issuing a deprecation warning by default for all users. Thus the configure default is changed to disable it, requiring users to pass --enable-sheepdog to build the driver.

Quoted 6 lines
---
 block/sheepdog.c           | 15 +++++++++++++++
 configure                  |  5 +++--
 docs/system/deprecated.rst |  9 +++++++++
 3 files changed, 27 insertions(+), 2 deletions(-)
Reviewed-by on v2 2/2
#6 lore raw

Re: [PATCH v2 2/2] block: deprecate the sheepdog block driver

In reply to Daniel P. Berrangé
Quoted 27 lines · Daniel P. Berrangé

This thread from a little over a year ago:

http://lists.wpkg.org/pipermail/sheepdog/2019-March/thread.html

states that sheepdog is no longer actively developed. The only mentioned users are some companies who are said to have it for legacy reasons with plans to replace it by Ceph. There is talk about cutting out existing features to turn it into a simple demo of how to write a distributed block service. There is no evidence of anyone working on that idea:

https://github.com/sheepdog/sheepdog/commits/master

No real commits to git since Jan 2018, and before then just some minor technical debt cleanup..

There is essentially no activity on the mailing list aside from patches to QEMU that get CC'd due to our MAINTAINERS entry.

Fedora packages for sheepdog failed to build from upstream source because of the more strict linker that no longer merges duplicate global symbols. Fedora patches it to add the missing "extern" annotations and presumably other distros do to, but upstream source remains broken.

There is only basic compile testing, no functional testing of the driver.

Quoted 7 lines

Since there are no build pre-requisites the sheepdog driver is currently enabled unconditionally. This would result in configure issuing a deprecation warning by default for all users. Thus the configure default is changed to disable it, requiring users to pass --enable-sheepdog to build the driver.

Signature · 2 lines
Best regards,
Vladimir
Reviewed-by on v2 2/2
#7 lore raw

Re: [PATCH v2 2/2] block: deprecate the sheepdog block driver

In reply to Daniel P. Berrangé
Quoted 45 lines · Daniel P. Berrangé

This thread from a little over a year ago:

http://lists.wpkg.org/pipermail/sheepdog/2019-March/thread.html

states that sheepdog is no longer actively developed. The only mentioned users are some companies who are said to have it for legacy reasons with plans to replace it by Ceph. There is talk about cutting out existing features to turn it into a simple demo of how to write a distributed block service. There is no evidence of anyone working on that idea:

https://github.com/sheepdog/sheepdog/commits/master

No real commits to git since Jan 2018, and before then just some minor technical debt cleanup..

There is essentially no activity on the mailing list aside from patches to QEMU that get CC'd due to our MAINTAINERS entry.

Fedora packages for sheepdog failed to build from upstream source because of the more strict linker that no longer merges duplicate global symbols. Fedora patches it to add the missing "extern" annotations and presumably other distros do to, but upstream source remains broken.

There is only basic compile testing, no functional testing of the driver.

Since there are no build pre-requisites the sheepdog driver is currently enabled unconditionally. This would result in configure issuing a deprecation warning by default for all users. Thus the configure default is changed to disable it, requiring users to pass --enable-sheepdog to build the driver.

---
 block/sheepdog.c           | 15 +++++++++++++++
 configure                  |  5 +++--
 docs/system/deprecated.rst |  9 +++++++++
 3 files changed, 27 insertions(+), 2 deletions(-)
block/sheepdog.c +0 -0
index cbbebc1aaf..7f68bd6a1a 100644

block/sheepdog.c typedef struct SheepdogInode { · line 247

Quoted 6 lines
block/sheepdog.c +3 -0
@@ -242,6 +242,17 @@ typedef struct SheepdogInode {
  */
 #define FNV1A_64_INIT ((uint64_t)0xcbf29ce484222325ULL)
 
+static void deprecation_warning(void)
+{
+    static bool warned = false;

Checkpatch warns there is no need to initialize static bool to false, otherwise:

Quoted 7 lines
block/sheepdog.c +7 -0
+
+    if (!warned) {
+        warn_report("the sheepdog block driver is deprecated and will be "
+                    "removed in a future release");
+        warned = true;
+    }
+}

[...]

Reviewed-by on v2 2/2
#8 lore raw

Re: [PATCH v2 2/2] block: deprecate the sheepdog block driver

In reply to Daniel P. Berrangé
Quoted 10 lines · Daniel P. Berrangé

This thread from a little over a year ago:

http://lists.wpkg.org/pipermail/sheepdog/2019-March/thread.html

states that sheepdog is no longer actively developed. The only mentioned users are some companies who are said to have it for legacy reasons with plans to replace it by Ceph. There is talk about cutting out existing features to turn it into a simple demo of how to write a distributed block service. There is no evidence of anyone working on that idea:

Quoted 4 lines

https://github.com/sheepdog/sheepdog/commits/master

No real commits to git since Jan 2018, and before then just some minor technical debt cleanup..

Drop the extra period.

Quoted 31 lines

There is essentially no activity on the mailing list aside from patches to QEMU that get CC'd due to our MAINTAINERS entry.

Fedora packages for sheepdog failed to build from upstream source because of the more strict linker that no longer merges duplicate global symbols. Fedora patches it to add the missing "extern" annotations and presumably other distros do to, but upstream source remains broken.

There is only basic compile testing, no functional testing of the driver.

Since there are no build pre-requisites the sheepdog driver is currently enabled unconditionally. This would result in configure issuing a deprecation warning by default for all users. Thus the configure default is changed to disable it, requiring users to pass --enable-sheepdog to build the driver.

---
 block/sheepdog.c           | 15 +++++++++++++++
 configure                  |  5 +++--
 docs/system/deprecated.rst |  9 +++++++++
 3 files changed, 27 insertions(+), 2 deletions(-)
block/sheepdog.c +0 -0
index cbbebc1aaf..7f68bd6a1a 100644

block/sheepdog.c typedef struct SheepdogInode { · line 247

Quoted 6 lines
block/sheepdog.c +3 -0
@@ -242,6 +242,17 @@ typedef struct SheepdogInode {
  */
 #define FNV1A_64_INIT ((uint64_t)0xcbf29ce484222325ULL)
 
+static void deprecation_warning(void)
+{
+    static bool warned = false;

Obey checkpatch :)

block/sheepdog.c typedef struct SheepdogInode { · line 251

Quoted 4 lines
block/sheepdog.c +4 -0
+
+    if (!warned) {
+        warn_report("the sheepdog block driver is deprecated and will be "
+                    "removed in a future release");

Similar warnings elsewhere don't say "will be removed".

Some of them are nice enough to advise what to use instead, but that may not be practical here.

Quoted 64 lines
block/sheepdog.c +8 -0
+        warned = true;
+    }
+}
+
 /*
  * 64 bit Fowler/Noll/Vo FNV-1a hash code
  */
@@ -1548,6 +1559,8 @@ static int sd_open(BlockDriverState *bs, QDict *options, int flags,
     char *buf = NULL;
     QemuOpts *opts;
 
+    deprecation_warning();
+
     s->bs = bs;
     s->aio_context = bdrv_get_aio_context(bs);
 
@@ -2007,6 +2020,8 @@ static int sd_co_create(BlockdevCreateOptions *options, Error **errp)
 
     assert(options->driver == BLOCKDEV_DRIVER_SHEEPDOG);
 
+    deprecation_warning();
+
     s = g_new0(BDRVSheepdogState, 1);
 
     /* Steal SocketAddress from QAPI, set NULL to prevent double free */
configure +3 -2
index 7564479008..c6af83f2e6 100755
@@ -533,7 +533,7 @@ vdi="yes"
 vvfat="yes"
 qed="yes"
 parallels="yes"
-sheepdog="yes"
+sheepdog="no"
 libxml2=""
 debug_mutex="no"
 libpmem=""
@@ -1941,7 +1941,7 @@ disabled with --disable-FEATURE, default is enabled if available:
   vvfat           vvfat image format support
   qed             qed image format support
   parallels       parallels image format support
-  sheepdog        sheepdog block driver support
+  sheepdog        sheepdog block driver support (deprecated)
   crypto-afalg    Linux AF_ALG crypto backend driver
   capstone        capstone disassembler support
   debug-mutex     mutex debugging support
@@ -7350,6 +7350,7 @@ if test "$parallels" = "yes" ; then
   echo "CONFIG_PARALLELS=y" >> $config_host_mak
 fi
 if test "$sheepdog" = "yes" ; then
+  add_to deprecated_features "sheepdog"
   echo "CONFIG_SHEEPDOG=y" >> $config_host_mak
 fi
 if test "$pty_h" = "yes" ; then
docs/system/deprecated.rst +1 -0
index 0cb8b01424..49b9f4b02e 100644
@@ -405,6 +405,15 @@ The above, converted to the current supported format::
 
   json:{"file.driver":"rbd", "file.pool":"rbd", "file.image":"name"}
 
+``sheepdog`` driver (since 5.2.0)

docs/system/deprecated.rst The above, converted to the current supported format:: · line 418

Quoted 10 lines
docs/system/deprecated.rst +8 -0
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The ``sheepdog`` block device driver is deprecated. The corresponding upstream
+server project is no longer actively maintained. Users are recommended to switch
+to an alternative distributed block device driver such as RBD. The
+``qemu-img convert`` command can be used to liberate existing data by moving
+it out of sheepdog volumes into an alternative storage backend.
+
 linux-user mode CPUs
 --------------------

Preferably with my nits addressed:

Reviewed-by on the v2 series
#9 lore raw
In reply to Daniel P. Berrangé
Quoted 15 lines · Daniel P. Berrangé

2 years back I proposed dropping the sheepdog mailing list from the MAINTAINERS file, but somehow the patch never got picked up:

https://lists.gnu.org/archive/html/qemu-block/2018-03/msg01048.html

So here I am with the same patch again.

This time I go further and deprecate the sheepdog driver entirely. See the rationale in the second patch commit message.

Daniel P. Berrangé (2):
block: drop moderated sheepdog mailing list from MAINTAINERS file
block: deprecate the sheepdog block driver

Quoted 10 lines
 MAINTAINERS                |  1 -
 block/sheepdog.c           | 15 +++++++++++++++
 configure                  |  5 +++--
 docs/system/deprecated.rst |  9 +++++++++
 4 files changed, 27 insertions(+), 3 deletions(-)

--
2.26.2

I don't know of anyone shipping this other than Fedora, and I certainly don't use it there.

Upstream looks like it's unmaintained now, with no commits in over two
years: https://github.com/sheepdog/sheepdog/commits/master

Can we also get a corresponding change to eliminate this from libvirt?

Signature · 1 line
真実はいつも一つ!/ Always, there's only one truth!
#10 lore raw
Quoted 27 lines · Neal Gompa
Quoted 15 lines · Daniel P. Berrangé

2 years back I proposed dropping the sheepdog mailing list from the MAINTAINERS file, but somehow the patch never got picked up:

https://lists.gnu.org/archive/html/qemu-block/2018-03/msg01048.html

So here I am with the same patch again.

This time I go further and deprecate the sheepdog driver entirely. See the rationale in the second patch commit message.

Daniel P. Berrangé (2):
block: drop moderated sheepdog mailing list from MAINTAINERS file
block: deprecate the sheepdog block driver

Quoted 10 lines
 MAINTAINERS                |  1 -
 block/sheepdog.c           | 15 +++++++++++++++
 configure                  |  5 +++--
 docs/system/deprecated.rst |  9 +++++++++
 4 files changed, 27 insertions(+), 3 deletions(-)

--
2.26.2

Quoted 7 lines

I don't know of anyone shipping this other than Fedora, and I certainly don't use it there.

Upstream looks like it's unmaintained now, with no commits in over two
years: https://github.com/sheepdog/sheepdog/commits/master

Can we also get a corresponding change to eliminate this from libvirt?

This is only deprecation in QEMU, the feature still exists and is intended to be as fully functional as in previous releases.

Assuming QEMU actually deletes the feature at end of the deprecation cycle, then libvirt would look at removing its own support.

Regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#11 lore raw
Quoted 35 lines · Daniel P. Berrangé
Quoted 33 lines · Neal Gompa
Quoted 15 lines · Daniel P. Berrangé

2 years back I proposed dropping the sheepdog mailing list from the MAINTAINERS file, but somehow the patch never got picked up:

https://lists.gnu.org/archive/html/qemu-block/2018-03/msg01048.html

So here I am with the same patch again.

This time I go further and deprecate the sheepdog driver entirely. See the rationale in the second patch commit message.

Daniel P. Berrangé (2):
block: drop moderated sheepdog mailing list from MAINTAINERS file
block: deprecate the sheepdog block driver

Quoted 10 lines
 MAINTAINERS                |  1 -
 block/sheepdog.c           | 15 +++++++++++++++
 configure                  |  5 +++--
 docs/system/deprecated.rst |  9 +++++++++
 4 files changed, 27 insertions(+), 3 deletions(-)

--
2.26.2

I don't know of anyone shipping this other than Fedora, and I certainly don't use it there.

Upstream looks like it's unmaintained now, with no commits in over two
years: https://github.com/sheepdog/sheepdog/commits/master

Quoted 8 lines
Quoted 1 line

Can we also get a corresponding change to eliminate this from libvirt?

This is only deprecation in QEMU, the feature still exists and is intended to be as fully functional as in previous releases.

Assuming QEMU actually deletes the feature at end of the deprecation cycle, then libvirt would look at removing its own support.

Does that stop us from deprecating it in libvirt though?

Signature · 1 line
真実はいつも一つ!/ Always, there's only one truth!
#12 lore raw
Quoted 38 lines · Neal Gompa
Quoted 37 lines · Daniel P. Berrangé
Quoted 34 lines · Neal Gompa
Quoted 15 lines · Daniel P. Berrangé

2 years back I proposed dropping the sheepdog mailing list from the MAINTAINERS file, but somehow the patch never got picked up:

   https://lists.gnu.org/archive/html/qemu-block/2018-03/msg01048.html

So here I am with the same patch again.

This time I go further and deprecate the sheepdog driver entirely. See the rationale in the second patch commit message.

Daniel P. Berrangé (2):
block: drop moderated sheepdog mailing list from MAINTAINERS file
block: deprecate the sheepdog block driver

Quoted 10 lines
  MAINTAINERS                |  1 -
  block/sheepdog.c           | 15 +++++++++++++++
  configure                  |  5 +++--
  docs/system/deprecated.rst |  9 +++++++++
  4 files changed, 27 insertions(+), 3 deletions(-)

--
2.26.2

I don't know of anyone shipping this other than Fedora, and I certainly don't use it there.

Upstream looks like it's unmaintained now, with no commits in over two
years: https://github.com/sheepdog/sheepdog/commits/master

Can we also get a corresponding change to eliminate this from libvirt?

Quoted 9 lines
Quoted 6 lines

This is only deprecation in QEMU, the feature still exists and is intended to be as fully functional as in previous releases.

Assuming QEMU actually deletes the feature at end of the deprecation cycle, then libvirt would look at removing its own support.

Does that stop us from deprecating it in libvirt though?

I think not. Libvirt is not intended to support all Qemu features and I'm sure it doesn't. Amd it can safely deprecate features even if they are not-deprecated in Qemu.

The only possible conflict is when Qemu wants to deprecate something that Libvirt wants to continue support for (or even can't live without).

Signature · 2 lines
Best regards,
Vladimir
#13 lore raw
In reply to Daniel P. Berrangé
Quoted 34 lines · Daniel Berrange
Quoted 33 lines · Neal Gompa
Quoted 15 lines · Daniel P. Berrangé

2 years back I proposed dropping the sheepdog mailing list from the MAINTAINERS file, but somehow the patch never got picked up:

https://lists.gnu.org/archive/html/qemu-block/2018-03/msg01048.html

So here I am with the same patch again.

This time I go further and deprecate the sheepdog driver entirely. See the rationale in the second patch commit message.

Daniel P. Berrangé (2):
block: drop moderated sheepdog mailing list from MAINTAINERS file
block: deprecate the sheepdog block driver

Quoted 10 lines
 MAINTAINERS                |  1 -
 block/sheepdog.c           | 15 +++++++++++++++
 configure                  |  5 +++--
 docs/system/deprecated.rst |  9 +++++++++
 4 files changed, 27 insertions(+), 3 deletions(-)

--
2.26.2

I don't know of anyone shipping this other than Fedora, and I certainly don't use it there.

Upstream looks like it's unmaintained now, with no commits in over two
years: https://github.com/sheepdog/sheepdog/commits/master

Quoted 7 lines
Quoted 1 line

Can we also get a corresponding change to eliminate this from libvirt?

This is only deprecation in QEMU, the feature still exists and is intended to be as fully functional as in previous releases.

Assuming QEMU actually deletes the feature at end of the deprecation cycle, then libvirt would look at removing its own support.

There are two sheepdog-related bits in libvirt which are IMO completely separate:

1) the blockdev backend handling for the 'sheepdog' protocol driver

This is the one connected to qemu's deprecation, but until there is a qemu version where sheepdog wasn't deprecated/removed supported by libvirt we can't really do much about it.

On the other hand it's just a few generators of arguments for -drive/-blockdev so the burden is very low.

2) the sheepdog storage driver

This one is completely separate from qemu and we can decide to deprecate/delete it at our own schedule as it will not influence the ability to start VMs.

The last non-housekeeping commit in that driver seems to be dated:
Thu Jun 18 16:20:42 2015

Similarly the burden of keeping this around is low, but I would not bat an eye if we remove it right away even.

#14 lore raw
In reply to Daniel P. Berrangé
Patchew URL: https://patchew.org/QEMU/[email protected]/

Patchew URL: https://patchew.org/QEMU/[email protected]/

Hi,

This series seems to have some coding style problems. See output below for more information:

Type: series
Message-id: [email protected]
Subject: [PATCH v2 0/2] block: deprecate the sheepdog driver

=== TEST SCRIPT BEGIN ===
#!/bin/bash
git rev-parse base > /dev/null || exit 0
git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram
./scripts/checkpatch.pl --mailback base..
=== TEST SCRIPT END ===

Switched to a new branch 'test'
2c11635 block: deprecate the sheepdog block driver
aa3d54c block: drop moderated sheepdog mailing list from MAINTAINERS file

=== OUTPUT BEGIN ===
1/2 Checking commit aa3d54cfa28f (block: drop moderated sheepdog mailing list from MAINTAINERS file) 2/2 Checking commit 2c1163558579 (block: deprecate the sheepdog block driver)
ERROR: do not initialise statics to 0 or NULL
#54: FILE: block/sheepdog.c:247:
+ static bool warned = false;

total: 1 errors, 0 warnings, 71 lines checked

Patch 2/2 has style problems, please review. If any of these errors are false positives report them to the maintainer, see CHECKPATCH in MAINTAINERS.

=== OUTPUT END ===

Test command exited with code: 1

The full log is available at
http://patchew.org/logs/[email protected]/testing.checkpatch/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to [email protected]