Update detect-raspbian.diff

Manual merge of version 14.2.21-1.1+rpi1 and 16.2.7+ds-4 to produce 16.2.7+ds-4+rpi1

Merge ceph (16.2.7+ds-4) import into refs/heads/workingbranch

Fix build with fmtlib 8.1.1

Bug: https://tracker.ceph.com/issues/53820
Bug-Debian: https://bugs.debian.org/1003469
Forwarded: no
Last-Update: 2022-01-14

+ changes in segment_manager.cc and segment_manager.h are backported from
  part of the large changes in https://github.com/ceph/ceph/commit/d5b0cd13
+ change in node_extent_accessor.h is not forwarded to upstream since it's
  a workaround. However it doesn't harm since it's just a error message
  which shouldn't happen anyway.

Gbp-Pq: Name fix-build-with-fmt-8.1.1.patch

Only yield under ARMv7 and above (#1176)

Origin: upstream, https://github.com/facebook/folly/commit/62d8e6e0b91ebd6f878f3066cd9b6e5f3c18a97b.patch
Last-Update: 2021-11-24

Gbp-Pq: Name only-yied-under-armv7-and-above.patch

cmake: test for 16-byte atomic support on mips also

Origin: upstream, https://github.com/ceph/ceph/commit/709a77f22010f03aee4a4c0ab930588944cb4a58
Last-Update: 2021-11-24

it's reported that a mips64el build host is able to pass the test of
CheckCxxAtomic without linking against libatomic, while librbd.so
fails to link due to failures like

/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_store_16'
/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_load_16'
/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_compare_exchange_16'

so we have to check the existence of __atomic_load_16 instruction on
mips architecture.

Gbp-Pq: Name cmake-test-for-16-bytes-atomic-support-on-mips-also.patch

[PATCH] src/isa-l/erasure_code: Fix text relocation on aarch64

Here is the bug report on ceph. https://tracker.ceph.com/issues/48681

Signed-off-by: luo rixin <luorixin@huawei.com>
Gbp-Pq: Name bug1917414.patch

[PATCH] rgw/radosgw-admin clarify error when email address already in use

The error message if you try and create an S3 user with an email
address that is already associated with another S3 account is very
confusing; this patch makes it much clearer

To reproduce:

radosgw-admin user create --uid=foo --display-name="Foo test" --email=bar@domain.invalid
radosgw-admin user create --uid=test --display-name="AN test" --email=bar@domain.invalid
could not create user: unable to parse parameters, user id mismatch, operation id: foo does not match: test

With this patch:

radosgw-admin user create --uid=test --display-name="AN test" --email=bar@domain.invalid
could not create user: unable to create user test because user id foo already exists with email bar@domain.invalid

Fixes: https://tracker.ceph.com/issues/49137
Fixes: https://tracker.ceph.com/issues/19411
Signed-off-by: Matthew Vernon <mv3@sanger.ac.uk>
(cherry picked from commit 05318d6f71e45a42a46518a0ef17047dfab83990)

Gbp-Pq: Name bug1914584.patch

[PATCH] arch,cmake: compile ppc.c on all powerpc machines

* cmake/modules/SIMDExt.cmake: define HAVE_PPC for 32-bit PowerPC.
* src/arch/CMakeLists.txt: compile ppc.c for all PowerPC architectures,
  including powerpc (32-bit PowerPC), ppc64el (64-bit Little Endian
  PowerPC) and ppc64 (64-bit Big Endian PowerPC).

before this change, ppc.c is only compiled if HAVE_POWER8 is defined.
but Power8 is a 64-bit PowerPC architecture. while in src/arch/probe.cc,
we check for `defined(__powerpc__) || defined(__ppc__)`, if this is
true, ceph_arch_ppc_probe() is used to check for the support of
Altivec. but on non-power8 PowerPC machines, the linker fails to find the
symbols like ceph_arch_ppc_probe(), as ppc.c is not compiled on them.

in this change, ppc.c is compiled on all PowerPC architectures, so that
ceph_arch_ppc_probe() is also available on non-power8 machines. this
change does not impact the behavior of non-power8 machines. because
on them, the runtime check would fail to detect the existence of
PPC_FEATURE2_VEC_CRYPTO instructions.

Reported-by: Mattias Ellert <mattias.ellert@physics.uu.se>
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
Gbp-Pq: Name compile-ppc.c-on-all-powerpc-machines.patch

Fix systemd ceph-osd.target

Forwarded: no
Last-Update: 2021-01-28

This helps when rebooting.

Gbp-Pq: Name fix-ceph-osd-systemd-target.patch

[PATCH] Remove logging every slow request details to monitors

 recent change https://tracker.ceph.com/issues/43975 was made to slow
 request logging to include detail on each operation in the cluster logs.
 With this change, detail for every slow request is always sent to the
 monitors and added to the cluster logs.

 This does not scale. Large, high-throughput clusters can overwhelm
 their monitors with spurious logs in the event of a performance issue.
 Disrupting the monitors can then cause further instability in the
 cluster.

 This SRU reverts the cluster logging of every slow request the osd is
 processing.

 The slow request clog change was added in nautilus (14.2.10) and
 octopus (15.2.0).

Signed-off-by: Gerald Yang <gerald.yang@canonical.com>
Gbp-Pq: Name disable-log-slow-requests.patch

Link with -pthread instead of -lpthread to fix FTBFS on riscv64

Forwarded: no
Last-Update: 2020-03-01

Gbp-Pq: Name riscv64-link-pthread.patch

add-option-to-disable-ceph-dencoder

===================================================================

Gbp-Pq: Name add-option-to-disable-ceph-dencoder.patch

Misc fixes for 32 bit architecture builds.

Forwarded: no

Gbp-Pq: Name 32bit-fixes.patch

fix-bash-completion-location

Gbp-Pq: Name fix-bash-completion-location

debian-armel-armhf-buildflags

Gbp-Pq: Name debian-armel-armhf-buildflags.patch

Adds max_connections to test display.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/3b8eb36676f70d06f8918ccf62029207c49cdda0
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable_test.patch

Makes SOMAXCONN user-configurable.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/febab7dc38c9671577603425c54c20f841e27f97
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable.patch

Adds max_connections to reference configuration.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/3b8eb36676f70d06f8918ccf62029207c49cdda0
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable_conf.patch

disable-crypto

===================================================================

Gbp-Pq: Name disable-crypto.patch

use --release 7 instead of -source/-target

Bug-Ubuntu: https://launchpad.net/bugs/1756854
Bug-Ubuntu: https://launchpad.net/bugs/1766998
Forwarded: no
Last-Update: 2018-04-24

Instead of -source/-target ceph should be build with --release for OpenJDK 9
or later so that the bootclasspath is also set, as per JEP-247, otherwise it
risks incurring into binary incompatibility when run with an earlier OpenJDK.
OpenJDK 11 minimum compatibility release has been updated to 7.
Last-Update: 2018-04-24
Gbp-Pq: Name update-java-source-target-flags.patch

This defines HAVE_REENTRANT_STRSIGNAL as sys_siglist no longer

Forwarded: no
Last-Update: 2020-09-21

 exists with glibc 2.32 and all programs should use strsignal instead.

Gbp-Pq: Name enable-strsignal.patch

ceph (16.2.7+ds-4) unstable; urgency=medium

  * Add fix-build-with-fmt-8.1.1.patch (Closes: #1003469).

[dgit import unpatched ceph 16.2.7+ds-4]

Import ceph_16.2.7+ds-4.debian.tar.xz

[dgit import tarball ceph 16.2.7+ds-4 ceph_16.2.7+ds-4.debian.tar.xz]

Merge ceph (16.2.7+ds-3) import into refs/heads/workingbranch

Only yield under ARMv7 and above (#1176)

Origin: upstream, https://github.com/facebook/folly/commit/62d8e6e0b91ebd6f878f3066cd9b6e5f3c18a97b.patch
Last-Update: 2021-11-24

Gbp-Pq: Name only-yied-under-armv7-and-above.patch

cmake: test for 16-byte atomic support on mips also

Origin: upstream, https://github.com/ceph/ceph/commit/709a77f22010f03aee4a4c0ab930588944cb4a58
Last-Update: 2021-11-24

it's reported that a mips64el build host is able to pass the test of
CheckCxxAtomic without linking against libatomic, while librbd.so
fails to link due to failures like

/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_store_16'
/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_load_16'
/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_compare_exchange_16'

so we have to check the existence of __atomic_load_16 instruction on
mips architecture.

Gbp-Pq: Name cmake-test-for-16-bytes-atomic-support-on-mips-also.patch

[PATCH] src/isa-l/erasure_code: Fix text relocation on aarch64

Here is the bug report on ceph. https://tracker.ceph.com/issues/48681

Signed-off-by: luo rixin <luorixin@huawei.com>
Gbp-Pq: Name bug1917414.patch

[PATCH] rgw/radosgw-admin clarify error when email address already in use

The error message if you try and create an S3 user with an email
address that is already associated with another S3 account is very
confusing; this patch makes it much clearer

To reproduce:

radosgw-admin user create --uid=foo --display-name="Foo test" --email=bar@domain.invalid
radosgw-admin user create --uid=test --display-name="AN test" --email=bar@domain.invalid
could not create user: unable to parse parameters, user id mismatch, operation id: foo does not match: test

With this patch:

radosgw-admin user create --uid=test --display-name="AN test" --email=bar@domain.invalid
could not create user: unable to create user test because user id foo already exists with email bar@domain.invalid

Fixes: https://tracker.ceph.com/issues/49137
Fixes: https://tracker.ceph.com/issues/19411
Signed-off-by: Matthew Vernon <mv3@sanger.ac.uk>
(cherry picked from commit 05318d6f71e45a42a46518a0ef17047dfab83990)

Gbp-Pq: Name bug1914584.patch

[PATCH] arch,cmake: compile ppc.c on all powerpc machines

* cmake/modules/SIMDExt.cmake: define HAVE_PPC for 32-bit PowerPC.
* src/arch/CMakeLists.txt: compile ppc.c for all PowerPC architectures,
  including powerpc (32-bit PowerPC), ppc64el (64-bit Little Endian
  PowerPC) and ppc64 (64-bit Big Endian PowerPC).

before this change, ppc.c is only compiled if HAVE_POWER8 is defined.
but Power8 is a 64-bit PowerPC architecture. while in src/arch/probe.cc,
we check for `defined(__powerpc__) || defined(__ppc__)`, if this is
true, ceph_arch_ppc_probe() is used to check for the support of
Altivec. but on non-power8 PowerPC machines, the linker fails to find the
symbols like ceph_arch_ppc_probe(), as ppc.c is not compiled on them.

in this change, ppc.c is compiled on all PowerPC architectures, so that
ceph_arch_ppc_probe() is also available on non-power8 machines. this
change does not impact the behavior of non-power8 machines. because
on them, the runtime check would fail to detect the existence of
PPC_FEATURE2_VEC_CRYPTO instructions.

Reported-by: Mattias Ellert <mattias.ellert@physics.uu.se>
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
Gbp-Pq: Name compile-ppc.c-on-all-powerpc-machines.patch

Fix systemd ceph-osd.target

Forwarded: no
Last-Update: 2021-01-28

This helps when rebooting.

Gbp-Pq: Name fix-ceph-osd-systemd-target.patch

[PATCH] Remove logging every slow request details to monitors

 recent change https://tracker.ceph.com/issues/43975 was made to slow
 request logging to include detail on each operation in the cluster logs.
 With this change, detail for every slow request is always sent to the
 monitors and added to the cluster logs.

 This does not scale. Large, high-throughput clusters can overwhelm
 their monitors with spurious logs in the event of a performance issue.
 Disrupting the monitors can then cause further instability in the
 cluster.

 This SRU reverts the cluster logging of every slow request the osd is
 processing.

 The slow request clog change was added in nautilus (14.2.10) and
 octopus (15.2.0).

Signed-off-by: Gerald Yang <gerald.yang@canonical.com>
Gbp-Pq: Name disable-log-slow-requests.patch

Link with -pthread instead of -lpthread to fix FTBFS on riscv64

Forwarded: no
Last-Update: 2020-03-01

Gbp-Pq: Name riscv64-link-pthread.patch

add-option-to-disable-ceph-dencoder

===================================================================

Gbp-Pq: Name add-option-to-disable-ceph-dencoder.patch

Misc fixes for 32 bit architecture builds.

Forwarded: no

Gbp-Pq: Name 32bit-fixes.patch

fix-bash-completion-location

Gbp-Pq: Name fix-bash-completion-location

debian-armel-armhf-buildflags

Gbp-Pq: Name debian-armel-armhf-buildflags.patch

Adds max_connections to test display.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/3b8eb36676f70d06f8918ccf62029207c49cdda0
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable_test.patch

Makes SOMAXCONN user-configurable.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/febab7dc38c9671577603425c54c20f841e27f97
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable.patch

Adds max_connections to reference configuration.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/3b8eb36676f70d06f8918ccf62029207c49cdda0
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable_conf.patch

disable-crypto

===================================================================

Gbp-Pq: Name disable-crypto.patch

use --release 7 instead of -source/-target

Bug-Ubuntu: https://launchpad.net/bugs/1756854
Bug-Ubuntu: https://launchpad.net/bugs/1766998
Forwarded: no
Last-Update: 2018-04-24

Instead of -source/-target ceph should be build with --release for OpenJDK 9
or later so that the bootclasspath is also set, as per JEP-247, otherwise it
risks incurring into binary incompatibility when run with an earlier OpenJDK.
OpenJDK 11 minimum compatibility release has been updated to 7.
Last-Update: 2018-04-24
Gbp-Pq: Name update-java-source-target-flags.patch

This defines HAVE_REENTRANT_STRSIGNAL as sys_siglist no longer

Forwarded: no
Last-Update: 2020-09-21

 exists with glibc 2.32 and all programs should use strsignal instead.

Gbp-Pq: Name enable-strsignal.patch

ceph (16.2.7+ds-3) unstable; urgency=medium

  * Rebuild source-only.
  * Fix cephadm user creation that was failing, and subsequent chown failing as
    the package installation as a consequence.

[dgit import unpatched ceph 16.2.7+ds-3]

Import ceph_16.2.7+ds-3.debian.tar.xz

[dgit import tarball ceph 16.2.7+ds-3 ceph_16.2.7+ds-3.debian.tar.xz]

Import ceph_16.2.7+ds.orig.tar.xz

[dgit import orig ceph_16.2.7+ds.orig.tar.xz]

Merge ceph (16.2.6+ds-11) import into refs/heads/workingbranch

Only yield under ARMv7 and above (#1176)

Origin: upstream, https://github.com/facebook/folly/commit/62d8e6e0b91ebd6f878f3066cd9b6e5f3c18a97b.patch
Last-Update: 2021-11-24

Gbp-Pq: Name only-yied-under-armv7-and-above.patch

cmake: test for 16-byte atomic support on mips also

Origin: upstream, https://github.com/ceph/ceph/commit/709a77f22010f03aee4a4c0ab930588944cb4a58
Last-Update: 2021-11-24

it's reported that a mips64el build host is able to pass the test of
CheckCxxAtomic without linking against libatomic, while librbd.so
fails to link due to failures like

/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_store_16'
/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_load_16'
/usr/bin/ld: ../../../lib/librbd.so.1.16.0: undefined reference to `__atomic_compare_exchange_16'

so we have to check the existence of __atomic_load_16 instruction on
mips architecture.

Gbp-Pq: Name cmake-test-for-16-bytes-atomic-support-on-mips-also.patch

[PATCH] src/isa-l/erasure_code: Fix text relocation on aarch64

Here is the bug report on ceph. https://tracker.ceph.com/issues/48681

Signed-off-by: luo rixin <luorixin@huawei.com>
Gbp-Pq: Name bug1917414.patch

[PATCH] rgw/radosgw-admin clarify error when email address already in use

The error message if you try and create an S3 user with an email
address that is already associated with another S3 account is very
confusing; this patch makes it much clearer

To reproduce:

radosgw-admin user create --uid=foo --display-name="Foo test" --email=bar@domain.invalid
radosgw-admin user create --uid=test --display-name="AN test" --email=bar@domain.invalid
could not create user: unable to parse parameters, user id mismatch, operation id: foo does not match: test

With this patch:

radosgw-admin user create --uid=test --display-name="AN test" --email=bar@domain.invalid
could not create user: unable to create user test because user id foo already exists with email bar@domain.invalid

Fixes: https://tracker.ceph.com/issues/49137
Fixes: https://tracker.ceph.com/issues/19411
Signed-off-by: Matthew Vernon <mv3@sanger.ac.uk>
(cherry picked from commit 05318d6f71e45a42a46518a0ef17047dfab83990)

Gbp-Pq: Name bug1914584.patch

[PATCH] arch,cmake: compile ppc.c on all powerpc machines

* cmake/modules/SIMDExt.cmake: define HAVE_PPC for 32-bit PowerPC.
* src/arch/CMakeLists.txt: compile ppc.c for all PowerPC architectures,
  including powerpc (32-bit PowerPC), ppc64el (64-bit Little Endian
  PowerPC) and ppc64 (64-bit Big Endian PowerPC).

before this change, ppc.c is only compiled if HAVE_POWER8 is defined.
but Power8 is a 64-bit PowerPC architecture. while in src/arch/probe.cc,
we check for `defined(__powerpc__) || defined(__ppc__)`, if this is
true, ceph_arch_ppc_probe() is used to check for the support of
Altivec. but on non-power8 PowerPC machines, the linker fails to find the
symbols like ceph_arch_ppc_probe(), as ppc.c is not compiled on them.

in this change, ppc.c is compiled on all PowerPC architectures, so that
ceph_arch_ppc_probe() is also available on non-power8 machines. this
change does not impact the behavior of non-power8 machines. because
on them, the runtime check would fail to detect the existence of
PPC_FEATURE2_VEC_CRYPTO instructions.

Reported-by: Mattias Ellert <mattias.ellert@physics.uu.se>
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
Gbp-Pq: Name compile-ppc.c-on-all-powerpc-machines.patch

Fix systemd ceph-osd.target

Forwarded: no
Last-Update: 2021-01-28

This helps when rebooting.

Gbp-Pq: Name fix-ceph-osd-systemd-target.patch

[PATCH] Remove logging every slow request details to monitors

 recent change https://tracker.ceph.com/issues/43975 was made to slow
 request logging to include detail on each operation in the cluster logs.
 With this change, detail for every slow request is always sent to the
 monitors and added to the cluster logs.

 This does not scale. Large, high-throughput clusters can overwhelm
 their monitors with spurious logs in the event of a performance issue.
 Disrupting the monitors can then cause further instability in the
 cluster.

 This SRU reverts the cluster logging of every slow request the osd is
 processing.

 The slow request clog change was added in nautilus (14.2.10) and
 octopus (15.2.0).

Signed-off-by: Gerald Yang <gerald.yang@canonical.com>
Gbp-Pq: Name disable-log-slow-requests.patch

Link with -pthread instead of -lpthread to fix FTBFS on riscv64

Forwarded: no
Last-Update: 2020-03-01

Gbp-Pq: Name riscv64-link-pthread.patch

add-option-to-disable-ceph-dencoder

===================================================================

Gbp-Pq: Name add-option-to-disable-ceph-dencoder.patch

Misc fixes for 32 bit architecture builds.

Forwarded: no

Gbp-Pq: Name 32bit-fixes.patch

fix-bash-completion-location

Gbp-Pq: Name fix-bash-completion-location

debian-armel-armhf-buildflags

Gbp-Pq: Name debian-armel-armhf-buildflags.patch

Adds max_connections to test display.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/3b8eb36676f70d06f8918ccf62029207c49cdda0
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable_test.patch

Makes SOMAXCONN user-configurable.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/febab7dc38c9671577603425c54c20f841e27f97
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable.patch

Adds max_connections to reference configuration.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/3b8eb36676f70d06f8918ccf62029207c49cdda0
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable_conf.patch

disable-crypto

===================================================================

Gbp-Pq: Name disable-crypto.patch

use --release 7 instead of -source/-target

Bug-Ubuntu: https://launchpad.net/bugs/1756854
Bug-Ubuntu: https://launchpad.net/bugs/1766998
Forwarded: no
Last-Update: 2018-04-24

Instead of -source/-target ceph should be build with --release for OpenJDK 9
or later so that the bootclasspath is also set, as per JEP-247, otherwise it
risks incurring into binary incompatibility when run with an earlier OpenJDK.
OpenJDK 11 minimum compatibility release has been updated to 7.
Last-Update: 2018-04-24
Gbp-Pq: Name update-java-source-target-flags.patch

This defines HAVE_REENTRANT_STRSIGNAL as sys_siglist no longer

Forwarded: no
Last-Update: 2020-09-21

 exists with glibc 2.32 and all programs should use strsignal instead.

Gbp-Pq: Name enable-strsignal.patch

ceph (16.2.6+ds-11) unstable; urgency=medium

  * Revert libcephfs2.symbols "fix".

[dgit import unpatched ceph 16.2.6+ds-11]

Import ceph_16.2.6+ds-11.debian.tar.xz

[dgit import tarball ceph 16.2.6+ds-11 ceph_16.2.6+ds-11.debian.tar.xz]

Import ceph_16.2.6+ds.orig.tar.xz

[dgit import orig ceph_16.2.6+ds.orig.tar.xz]

Merge version 14.2.21-1+rpi1 and 14.2.21-1.1 to produce 14.2.21-1.1+rpi1

Merge ceph (14.2.21-1.1) import into refs/heads/workingbranch

[PATCH] arch,cmake: compile ppc.c on all powerpc machines

* cmake/modules/SIMDExt.cmake: define HAVE_PPC for 32-bit PowerPC.
* src/arch/CMakeLists.txt: compile ppc.c for all PowerPC architectures,
  including powerpc (32-bit PowerPC), ppc64el (64-bit Little Endian
  PowerPC) and ppc64 (64-bit Big Endian PowerPC).

before this change, ppc.c is only compiled if HAVE_POWER8 is defined.
but Power8 is a 64-bit PowerPC architecture. while in src/arch/probe.cc,
we check for `defined(__powerpc__) || defined(__ppc__)`, if this is
true, ceph_arch_ppc_probe() is used to check for the support of
Altivec. but on non-power8 PowerPC machines, the linker fails to find the
symbols like ceph_arch_ppc_probe(), as ppc.c is not compiled on them.

in this change, ppc.c is compiled on all PowerPC architectures, so that
ceph_arch_ppc_probe() is also available on non-power8 machines. this
change does not impact the behavior of non-power8 machines. because
on them, the runtime check would fail to detect the existence of
PPC_FEATURE2_VEC_CRYPTO instructions.

Reported-by: Mattias Ellert <mattias.ellert@physics.uu.se>
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
Gbp-Pq: Name compile-ppc.c-on-all-powerpc-machines.patch

allow BGP-to-the-host style binding

Forwarded: no
Last-Update: 2021-04-21

Gbp-Pq: Name allow-bgp-to-host.patch

Fix systemd ceph-osd.target

Forwarded: no
Last-Update: 2021-01-28

This helps when rebooting.

Gbp-Pq: Name fix-ceph-osd-systemd-target.patch

Another cmakelists fix

Forwarded: no
Last-Update: 2021-01-08

This fixes the last Boost 1.74 compatibility problems.

Gbp-Pq: Name another-cmakelists-fix.patch

cmake: add 1.74 to known versions

Bug-Debian: https://bugs.debian.org/977243
Origin: upstream, https://github.com/ceph/ceph/commit/b6a94da6149e50bdd43752919d7c01b04c59f79e.patch
Last-Update: 2020-12-13

Gbp-Pq: Name cmake_add_1.74_to_known_versions.patch

cmake: define BOOST_ASIO_USE_TS_EXECUTOR_AS_DEFAULT for

Signed-off-by: Kefu Chai <kchai@redhat.com>
Origin: upstream, https://github.com/ceph/ceph/commit/3d708219092d0e89a1434c30ffc8a4999f062cc0.patch
Bug-Debian: https://bugs.debian.org/977243
Last-Update: 2021-03-24

Boost.Asio users

see also
https://www.boost.org/doc/libs/1_74_0/doc/html/boost_asio/std_executors.html#boost_asio.std_executors.polymorphic_i_o_executor

we could use `asio::any_io_executor` later on though for better
performance.

also, define CMP0093, so FindBoost reports Boost_VERSION in x.y.z
format. it is simpler to use `VERSION_GREATER_EQUAL` to compare its
version with 1.74 instead of its C macro version ("107000").

Gbp-Pq: Name cmake_define_BOOST_ASIO_USE_TS_EXECUTOR_AS_DEFAULT_for_Boost.Asio_users.patch

Make Ceph Python 3.9 aware

Forwarded: no
Last-Update: 2020-11-28

Add versions of interpreters Ceph didn't know about.

Gbp-Pq: Name make-ceph-python-3.9-aware.patch

mds-purgequeue-use_uint64_t

===================================================================

Gbp-Pq: Name mds-purgequeue-use_uint64_t.patch

Link with -pthread instead of -lpthread to fix FTBFS on riscv64

Forwarded: no
Last-Update: 2020-03-01

Gbp-Pq: Name riscv64-link-pthread.patch

add-option-to-disable-ceph-dencoder

===================================================================

Gbp-Pq: Name add-option-to-disable-ceph-dencoder.patch

fix-bash-completion-location

Gbp-Pq: Name fix-bash-completion-location

debian-armel-armhf-buildflags

Gbp-Pq: Name debian-armel-armhf-buildflags.patch

[PATCH] os/bluestore/BlueFS: use uint64_t for `len`

change the type of parameter `len` of `BlueFS::_read_random()` from
`size_t` to `uint64_t`.

i think the type of `size_t` comes from
`rocksdb::RandomAccessFile::Read(uint64_t offset, size_t n,
rocksdb::Slice* result, char* scratch)`. and when we implement this
method, we continued using `n`'s type. but, we are using it with
`std::min()`, for instance, where the template parameter type deduction
fails if the lhs and rhs parameters' types are different. so probaly the
better solution is to use `uint64_t` directly to avoid the the cast and
specializing the template.

Signed-off-by: Kefu Chai <kchai@redhat.com>
Gbp-Pq: Name bluefs-use-uint64_t-for-len.patch

Adds max_connections to test display.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/3b8eb36676f70d06f8918ccf62029207c49cdda0
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable_test.patch

Makes SOMAXCONN user-configurable.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/febab7dc38c9671577603425c54c20f841e27f97
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable.patch

Adds max_connections to reference configuration.

Origin: upstream, https://github.com/civetweb/civetweb/pull/776/commits/3b8eb36676f70d06f8918ccf62029207c49cdda0
Bug: https://github.com/civetweb/civetweb/issues/775
Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109

Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/ceph/+bug/1838109
Gbp-Pq: Name civetweb-755-1.8-somaxconn-configurable_conf.patch

Avoid use of size_t when necessary

Forwarded: no

On 32 bit architectures size_t is not a 64 bit type, which
causes comparison mismatch failures during compilation.

Gbp-Pq: Name 32bit-avoid-size_t.patch

Avoid overloading on 32 bit architectures

Forwarded: no

unsigned and size_t are equivalent on 32 bit architectures,
so only define the size_t based overload of advance on 64
bit architectures.
https://wiki.debian.org/ArchitectureSpecificsMemo

Gbp-Pq: Name 32bit-avoid-overloading.patch

disable-crypto

===================================================================

Gbp-Pq: Name disable-crypto.patch

use --release 7 instead of -source/-target

Bug-Ubuntu: https://launchpad.net/bugs/1756854
Bug-Ubuntu: https://launchpad.net/bugs/1766998
Forwarded: no
Last-Update: 2018-04-24

Instead of -source/-target ceph should be build with --release for OpenJDK 9
or later so that the bootclasspath is also set, as per JEP-247, otherwise it
risks incurring into binary incompatibility when run with an earlier OpenJDK.
OpenJDK 11 minimum compatibility release has been updated to 7.
Last-Update: 2018-04-24
Gbp-Pq: Name update-java-source-target-flags.patch

ceph (14.2.21-1.1) unstable; urgency=medium

  * Non-maintainer upload.
  * Compile ppc.c on all powerpc machines. Fixes 32 bit powerpc build
    (Closes: #993642).

[dgit import unpatched ceph 14.2.21-1.1]

Import ceph_14.2.21-1.1.debian.tar.xz

[dgit import tarball ceph 14.2.21-1.1 ceph_14.2.21-1.1.debian.tar.xz]

Merge version 14.2.20-2+rpi1 and 14.2.21-1 to produce 14.2.21-1+rpi1

Merge ceph (14.2.21-1) import into refs/heads/workingbranch

Import ceph_14.2.21.orig.tar.gz

[dgit import orig ceph_14.2.21.orig.tar.gz]

allow BGP-to-the-host style binding

Forwarded: no
Last-Update: 2021-04-21

Gbp-Pq: Name allow-bgp-to-host.patch

Fix systemd ceph-osd.target

Forwarded: no
Last-Update: 2021-01-28

This helps when rebooting.

Gbp-Pq: Name fix-ceph-osd-systemd-target.patch

Another cmakelists fix

Forwarded: no
Last-Update: 2021-01-08

This fixes the last Boost 1.74 compatibility problems.

Gbp-Pq: Name another-cmakelists-fix.patch

cmake: add 1.74 to known versions

Bug-Debian: https://bugs.debian.org/977243
Origin: upstream, https://github.com/ceph/ceph/commit/b6a94da6149e50bdd43752919d7c01b04c59f79e.patch
Last-Update: 2020-12-13

Gbp-Pq: Name cmake_add_1.74_to_known_versions.patch

cmake: define BOOST_ASIO_USE_TS_EXECUTOR_AS_DEFAULT for

Signed-off-by: Kefu Chai <kchai@redhat.com>
Origin: upstream, https://github.com/ceph/ceph/commit/3d708219092d0e89a1434c30ffc8a4999f062cc0.patch
Bug-Debian: https://bugs.debian.org/977243
Last-Update: 2021-03-24

Boost.Asio users

see also
https://www.boost.org/doc/libs/1_74_0/doc/html/boost_asio/std_executors.html#boost_asio.std_executors.polymorphic_i_o_executor

we could use `asio::any_io_executor` later on though for better
performance.

also, define CMP0093, so FindBoost reports Boost_VERSION in x.y.z
format. it is simpler to use `VERSION_GREATER_EQUAL` to compare its
version with 1.74 instead of its C macro version ("107000").

Gbp-Pq: Name cmake_define_BOOST_ASIO_USE_TS_EXECUTOR_AS_DEFAULT_for_Boost.Asio_users.patch