[PATCH rdma-core 0/6] redhat/spec: various fixes and cleanups

[PATCH rdma-core 0/6] redhat/spec: various fixes and cleanups

[Jarod Wilson]

This is a set of patches to fix up a few things in the Red Hat (Fedora)
rdma-core spec file that came up as part of the Fedora new package review
process, where we're working on getting rdma-core into Fedora in place of
the old split packages.

Fedora package review bug:
  https://bugzilla.redhat.com/show_bug.cgi?id=1404043

Jarod Wilson (6):
  redhat/spec: call it srp_daemon in desc, not srptools
  redhat/spec: improve base package description
  redhat/spec: use fully versioned Requires on base package
  redhat/spec: fix up %post scriptlets
  redhat/spec: clean up rsocket packaging
  redhat/spec: make sure docdir is owned

 redhat/rdma-core.spec | 65 +++++++++++++++++++++++++++++++++++++--------------
 1 file changed, 47 insertions(+), 18 deletions(-)

-- 
2.10.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH rdma-core 1/6] redhat/spec: call it srp_daemon in desc, not srptools

[Jarod Wilson]

Signed-off-by: Jarod Wilson <jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 redhat/rdma-core.spec | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/redhat/rdma-core.spec b/redhat/rdma-core.spec
index 801bb94..82cd149 100644
--- a/redhat/rdma-core.spec
+++ b/redhat/rdma-core.spec
@@ -187,7 +187,7 @@ Requires(postun): systemd-units
 Requires: rdma-core
 
 %description -n srp_daemon
-In conjunction with the kernel ib_srp driver, srptools allows you to
+In conjunction with the kernel ib_srp driver, srp_daemon allows you to
 discover and use SCSI devices via the SCSI RDMA Protocol over InfiniBand.
 
 %prep
-- 
2.10.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH rdma-core 2/6] redhat/spec: improve base package description

[Jarod Wilson]

It was a bit lacking, add some meat to it.

Signed-off-by: Jarod Wilson <jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 redhat/rdma-core.spec | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/redhat/rdma-core.spec b/redhat/rdma-core.spec
index 82cd149..f9ebad1 100644
--- a/redhat/rdma-core.spec
+++ b/redhat/rdma-core.spec
@@ -20,6 +20,7 @@ BuildRequires: pkgconfig(libnl-3.0)
 BuildRequires: pkgconfig(libnl-route-3.0)
 BuildRequires: valgrind-devel
 
+Requires: dracut
 # Red Hat/Fedora previously shipped redhat/ as a stand-alone
 # package called 'rdma', which we're supplanting here.
 Provides: rdma = %{version}-%{release}
@@ -41,7 +42,9 @@ BuildRequires: make
 %endif
 
 %description
-RDMA core userspace infrastructure and documentation.
+RDMA core userspace infrastructure and documentation, including initscripts,
+kernel driver-specific modprobe override configs, IPoIB network scripts,
+dracut rules, and the rdma-ndd utility.
 
 %package devel
 Summary: RDMA core development libraries and headers
-- 
2.10.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH rdma-core 3/6] redhat/spec: use fully versioned Requires on base package

[Jarod Wilson]

Fedora packaging guidelines spell this at as an explicit requirement.

Signed-off-by: Jarod Wilson <jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 redhat/rdma-core.spec | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/redhat/rdma-core.spec b/redhat/rdma-core.spec
index f9ebad1..42d624b 100644
--- a/redhat/rdma-core.spec
+++ b/redhat/rdma-core.spec
@@ -66,7 +66,7 @@ RDMA core development libraries and headers.
 Summary: A library and drivers for direct userspace use of RDMA (InfiniBand/iWARP) hardware
 Requires(post): /sbin/ldconfig
 Requires(postun): /sbin/ldconfig
-Requires: rdma-core
+Requires: %{name}%{?_isa} = %{version}-%{release}
 Provides: libcxgb3 = %{version}-%{release}
 Obsoletes: libcxgb3 < %{version}-%{release}
 Provides: libcxgb4 = %{version}-%{release}
@@ -124,7 +124,7 @@ Summary: InfiniBand Communication Manager Assistant
 Requires(post): systemd-units
 Requires(preun): systemd-units
 Requires(postun): systemd-units
-Requires: rdma-core
+Requires: %{name}%{?_isa} = %{version}-%{release}
 
 %description -n ibacm
 The ibacm daemon helps reduce the load of managing path record lookups on
@@ -141,7 +141,7 @@ Summary: iWarp Port Mapper userspace daemon
 Requires(post): systemd-units
 Requires(preun): systemd-units
 Requires(postun): systemd-units
-Requires: rdma-core
+Requires: %{name}%{?_isa} = %{version}-%{release}
 
 %description -n iwpmd
 iwpmd provides a userspace service for iWarp drivers to claim
@@ -150,7 +150,7 @@ tcp ports through the standard socket interface.
 %package -n libibcm
 Summary: Userspace InfiniBand Connection Manager
 ExcludeArch: s390 s390x
-Requires: rdma-core
+Requires: %{name}%{?_isa} = %{version}-%{release}
 
 %description -n libibcm
 libibcm provides a userspace library that handles the majority of the low
@@ -158,7 +158,7 @@ level work required to open an RDMA connection between two machines.
 
 %package -n libibumad
 Summary: OpenFabrics Alliance InfiniBand umad (userspace management datagram) library
-Requires: rdma-core
+Requires: %{name}%{?_isa} = %{version}-%{release}
 
 %description -n libibumad
 libibumad provides the userspace management datagram (umad) library
@@ -167,7 +167,7 @@ are used by the IB diagnostic and management tools, including OpenSM.
 
 %package -n librdmacm
 Summary: Userspace RDMA Connection Manager
-Requires: rdma-core
+Requires: %{name}%{?_isa} = %{version}-%{release}
 
 %description -n librdmacm
 librdmacm provides a userspace RDMA Communication Managment API.
@@ -187,7 +187,7 @@ Obsoletes: openib-srptools <= 0.0.6
 Requires(post): systemd-units
 Requires(preun): systemd-units
 Requires(postun): systemd-units
-Requires: rdma-core
+Requires: %{name}%{?_isa} = %{version}-%{release}
 
 %description -n srp_daemon
 In conjunction with the kernel ib_srp driver, srp_daemon allows you to
-- 
2.10.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH rdma-core 4/6] redhat/spec: fix up %post scriptlets

[Jarod Wilson]

The split packaging spec is missing systemd restart bits for ibacm and
iwpmd, and has incorrect ldconfig calls. The base package has no libs in
it, that should have been on the libibverbs sub-package, and the other
lib sub-packages need one as well.

Signed-off-by: Jarod Wilson <jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 redhat/rdma-core.spec | 37 +++++++++++++++++++++++++++++++------
 1 file changed, 31 insertions(+), 6 deletions(-)

diff --git a/redhat/rdma-core.spec b/redhat/rdma-core.spec
index 42d624b..1f51720 100644
--- a/redhat/rdma-core.spec
+++ b/redhat/rdma-core.spec
@@ -267,20 +267,45 @@ install -D -m0644 redhat/srp_daemon.service %{buildroot}%{_unitdir}/
 # Delete the package's init.d scripts
 rm -rf %{buildroot}/%{_initrddir}/
 
-%post -p /sbin/ldconfig
-%postun -p /sbin/ldconfig
+# libibverbs
+%post -n libibverbs -p /sbin/ldconfig
+%postun -n libibverbs -p /sbin/ldconfig
 
+# libibcm
+%post -n libibcm -p /sbin/ldconfig
+%postun -n libibcm -p /sbin/ldconfig
+
+# libibumad
+%post -n libibumad -p /sbin/ldconfig
+%postun -n libibumad -p /sbin/ldconfig
+
+# librdmacm
+%post -n librdmacm -p /sbin/ldconfig
+%postun -n librdmacm -p /sbin/ldconfig
+
+# ibacm
 %post -n ibacm
 %systemd_post ibacm.service
-
 %preun -n ibacm
 %systemd_preun ibacm.service
-
 %postun -n ibacm
 %systemd_postun_with_restart ibacm.service
 
-%post -n libibcm -p /sbin/ldconfig
-%postun -n libibcm -p /sbin/ldconfig
+# srp_daemon
+%post -n srp_daemon
+%systemd_post srp_daemon.service
+%preun -n srp_daemon
+%systemd_preun srp_daemon.service
+%postun -n srp_daemon
+%systemd_postun_with_restart srp_daemon.service
+
+# iwpmd
+%post -n iwpmd
+%systemd_post iwpmd.service
+%preun -n iwpmd
+%systemd_preun iwpmd.service
+%postun -n iwpmd
+%systemd_postun_with_restart iwpmd.service
 
 %files
 %dir %{_sysconfdir}/rdma
-- 
2.10.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH rdma-core 5/6] redhat/spec: clean up rsocket packaging

[Jarod Wilson]

Nothing owned libdir/rsocket, and librdmacm was laying down dangling
symlinks to librspreload.so, so just put it all in librdmacm.

Signed-off-by: Jarod Wilson <jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 redhat/rdma-core.spec | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/redhat/rdma-core.spec b/redhat/rdma-core.spec
index 1f51720..96a67a4 100644
--- a/redhat/rdma-core.spec
+++ b/redhat/rdma-core.spec
@@ -336,13 +336,11 @@ rm -rf %{buildroot}/%{_initrddir}/
 %doc %{_docdir}/%{name}-%{version}/MAINTAINERS
 %{_includedir}/*
 %{_libdir}/lib*.so
-%{_libdir}/rsocket/*.so
 %{_mandir}/man3/ibv_*
 %{_mandir}/man3/rdma*
 %{_mandir}/man3/umad*
 %{_mandir}/man3/*_to_ibv_rate.*
 %{_mandir}/man7/rdma_cm.*
-%{_mandir}/man7/rsocket.*
 
 %files -n libibverbs
 %dir %{_sysconfdir}/libibverbs.d
@@ -389,8 +387,10 @@ rm -rf %{buildroot}/%{_initrddir}/
 
 %files -n librdmacm
 %{_libdir}/librdmacm*.so.*
-%{_libdir}/rsocket/*.so.*
+%dir %{_libdir}/rsocket
+%{_libdir}/rsocket/*.so
 %doc %{_docdir}/%{name}-%{version}/librdmacm.md
+%{_mandir}/man7/rsocket.*
 
 %files -n librdmacm-utils
 %{_bindir}/cmtime
-- 
2.10.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH rdma-core 6/6] redhat/spec: make sure docdir is owned

[Jarod Wilson]

Signed-off-by: Jarod Wilson <jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 redhat/rdma-core.spec | 1 +
 1 file changed, 1 insertion(+)

diff --git a/redhat/rdma-core.spec b/redhat/rdma-core.spec
index 96a67a4..3bb16ed 100644
--- a/redhat/rdma-core.spec
+++ b/redhat/rdma-core.spec
@@ -309,6 +309,7 @@ rm -rf %{buildroot}/%{_initrddir}/
 
 %files
 %dir %{_sysconfdir}/rdma
+%dir %{_docdir}/%{name}-%{version}
 %doc %{_docdir}/%{name}-%{version}/README.md
 %config(noreplace) %{_sysconfdir}/rdma/*
 %config(noreplace) %{_sysconfdir}/udev/rules.d/*
-- 
2.10.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-core 0/6] redhat/spec: various fixes and cleanups

[Jason Gunthorpe]

On Wed, Dec 21, 2016 at 05:08:34PM -0500, Jarod Wilson wrote:

> Fedora package review bug:
>   https://bugzilla.redhat.com/show_bug.cgi?id=1404043

To elaborate on that discussion..

>> lrwxrwxrwx. 1 honli honli 15 Dec 20 00:37 librspreload.so.1 -> librspreload.so
>> lrwxrwxrwx. 1 honli honli 15 Dec 20 00:37 librspreload.so.1.0.0 -> librspreload.so

> This looks like an upstream problem, just doing a build in a freshly
> unpacked tree sans-rpm has similar results. I'll address it there.

I see you found the packaging bug.. FYI the mechanism in cmake is to
produce build/lib/librspreload.so and no symlinks.

Two symlinks are produces under build/librdmacm/ but those are just
artifacts of how the install install process works for symlinks.

Once installed via DESTDIR=`pwd`/inst ninja install everything ends up
properly:

$ ls -l build/inst/usr/local/lib/rsocket/ 
-rw-r--r-- 1 jgg orc 117696 Dec 12 17:29 librspreload.so
lrwxrwxrwx 1 jgg orc     15 Dec 21 16:12 librspreload.so.1 -> librspreload.so
lrwxrwxrwx 1 jgg orc     15 Dec 21 16:12 librspreload.so.1.0.0 -> librspreload.so

It is appropriate for all these links to be in the same package as the
.so - these are not 'devel' links. They are compat for people using
the wrong name of the preload library. Eg users should be doing

LD_PRELOAD=/usr/lib/rsocket/librspreload.so foo

But since the old build system didn't build with the plugin option we
go out of our way to include these historical symlinks just incase
someone is using them improperly.

>> The service had been named as "srpd" since 2009, should we keep the old name?

> I'd stick with following upstream.

srp_daemon.service hasn't migrated out of redhat/ yet, I've got no
objection to adding a systemd alias for srpd if it is a well known
historical name..

+[Install]
+Alias=srpd.service

> Question 2: why libibacmp.so.* removed for new ibacm-12-1.fc26 pkg?
> (libibcm.f26 keeps libibcm.so.1.0.12)

The presence of the .so and .so.1 files was a bug in the old build
system that did not call libtool with the plugin option. This library
is only for dlopen and should never be used by the linker, so no need
for the aliases.

> Comment 2: iwpmd no longer start after syslog.target, as it does not write
> any log files. iwpmd send log message to /var/log/messages.

syslog.target is obsolete, Upstream systemd says not to use it.

> honli: libnl3-devel is required for iwpmd libibverbs

As you say, taken care of by pkgconfig(libnl-3.0), etc

> BuildRequires: valgrind-devel
> honli: systemd and dracut also needed.

Not to build?

> honli: Please add "Requires:" entries against rsyslog, systemd, kmod,
> honli: logrotate, initscripts, and dracut.

It might make sense to depend on kmod in rdma core since it is using
the module loading infrastructure..

rsyslog/logrotate are just examples for srp, as packaged it uses the
standard syslog logger, so they are not Requires

dracut.. we don't need it, but if it is present there are plugins
included, not a Requires..

Not sure why initscripts?

systemd is automatic via rpm, right?

> I could, but this situation is temporary, there will be an upstream
> release and a full URL shortly. This was noted in comment #1, but I
> could add that same text to

Use this for now:

https://github.com/linux-rdma/rdma-core/archive/v12-rc2.tar.gz

Best to make sure the release process is going to work :)

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-core 5/6] redhat/spec: clean up rsocket packaging

[Jarod Wilson]

On 2016-12-21 5:08 PM, Jarod Wilson wrote:
> Nothing owned libdir/rsocket, and librdmacm was laying down dangling
> symlinks to librspreload.so, so just put it all in librdmacm.
>
> Signed-off-by: Jarod Wilson <jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> ---
>  redhat/rdma-core.spec | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/redhat/rdma-core.spec b/redhat/rdma-core.spec
> index 1f51720..96a67a4 100644
> --- a/redhat/rdma-core.spec
> +++ b/redhat/rdma-core.spec
> @@ -336,13 +336,11 @@ rm -rf %{buildroot}/%{_initrddir}/
>  %doc %{_docdir}/%{name}-%{version}/MAINTAINERS
>  %{_includedir}/*
>  %{_libdir}/lib*.so
> -%{_libdir}/rsocket/*.so
>  %{_mandir}/man3/ibv_*
>  %{_mandir}/man3/rdma*
>  %{_mandir}/man3/umad*
>  %{_mandir}/man3/*_to_ibv_rate.*
>  %{_mandir}/man7/rdma_cm.*
> -%{_mandir}/man7/rsocket.*
>
>  %files -n libibverbs
>  %dir %{_sysconfdir}/libibverbs.d
> @@ -389,8 +387,10 @@ rm -rf %{buildroot}/%{_initrddir}/
>
>  %files -n librdmacm
>  %{_libdir}/librdmacm*.so.*
> -%{_libdir}/rsocket/*.so.*
> +%dir %{_libdir}/rsocket
> +%{_libdir}/rsocket/*.so

Minor error here, should be rsocket/*.so* to pick up the symlinks too, 
at least until they get neutered.

>  %doc %{_docdir}/%{name}-%{version}/librdmacm.md
> +%{_mandir}/man7/rsocket.*
>
>  %files -n librdmacm-utils
>  %{_bindir}/cmtime
>


-- 
Jarod Wilson
jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-core 0/6] redhat/spec: various fixes and cleanups

[Jarod Wilson]

On 2016-12-21 6:40 PM, Jason Gunthorpe wrote:
> On Wed, Dec 21, 2016 at 05:08:34PM -0500, Jarod Wilson wrote:
>
>> Fedora package review bug:
>>   https://bugzilla.redhat.com/show_bug.cgi?id=1404043
>
> To elaborate on that discussion..
>
>>> lrwxrwxrwx. 1 honli honli 15 Dec 20 00:37 librspreload.so.1 -> librspreload.so
>>> lrwxrwxrwx. 1 honli honli 15 Dec 20 00:37 librspreload.so.1.0.0 -> librspreload.so
>
>> This looks like an upstream problem, just doing a build in a freshly
>> unpacked tree sans-rpm has similar results. I'll address it there.
>
> I see you found the packaging bug.. FYI the mechanism in cmake is to
> produce build/lib/librspreload.so and no symlinks.
>
> Two symlinks are produces under build/librdmacm/ but those are just
> artifacts of how the install install process works for symlinks.
>
> Once installed via DESTDIR=`pwd`/inst ninja install everything ends up
> properly:
>
> $ ls -l build/inst/usr/local/lib/rsocket/
> -rw-r--r-- 1 jgg orc 117696 Dec 12 17:29 librspreload.so
> lrwxrwxrwx 1 jgg orc     15 Dec 21 16:12 librspreload.so.1 -> librspreload.so
> lrwxrwxrwx 1 jgg orc     15 Dec 21 16:12 librspreload.so.1.0.0 -> librspreload.so
>
> It is appropriate for all these links to be in the same package as the
> .so - these are not 'devel' links. They are compat for people using
> the wrong name of the preload library. Eg users should be doing

Yeah, they look correct in the resulting binaries after some tweaks.

>>> The service had been named as "srpd" since 2009, should we keep the old name?
>
>> I'd stick with following upstream.
>
> srp_daemon.service hasn't migrated out of redhat/ yet, I've got no
> objection to adding a systemd alias for srpd if it is a well known
> historical name..
>
> +[Install]
> +Alias=srpd.service

That works too.

>> BuildRequires: valgrind-devel
>> honli: systemd and dracut also needed.
>
> Not to build?

Argh. From https://fedoraproject.org/wiki/Packaging:Systemd

Packaging
Filesystem locations

Packages with systemd unit files must put them into %{_unitdir}. 
%{_unitdir} evaluates to /lib/systemd/system on all Fedora systems 
(F-15+). Unit files are architecture independent (hence, not %{_lib}) 
and needed early in the boot process.

Please note that in order for the %{_unitdir} macro to exist, your 
package must have:

BuildRequires: systemd

So it seems BR: systemd *is* necessary. Not so sure about dracut. 
Frankly, I think that's poor design, and the macros should be in one of 
the core rpm macros packages, but oh well.

I'm not aware of any dracut build deps, and indeed, my build in a Fedora 
rawhide mock chroot worked fine w/just systemd added.

>> honli: Please add "Requires:" entries against rsyslog, systemd, kmod,
>> honli: logrotate, initscripts, and dracut.
>
> It might make sense to depend on kmod in rdma core since it is using
> the module loading infrastructure..

The kernel itself already requires it, so unless you're installing on a 
system that somehow has no kernel installed, it's entirely redundant.

> rsyslog/logrotate are just examples for srp, as packaged it uses the
> standard syslog logger, so they are not Requires
>
> dracut.. we don't need it, but if it is present there are plugins
> included, not a Requires..

And dracut is another one already required by the kernel.

> Not sure why initscripts?
>
> systemd is automatic via rpm, right?

Also required by the kernel.

>> I could, but this situation is temporary, there will be an upstream
>> release and a full URL shortly. This was noted in comment #1, but I
>> could add that same text to
>
> Use this for now:
>
> https://github.com/linux-rdma/rdma-core/archive/v12-rc2.tar.gz
>
> Best to make sure the release process is going to work :)

Ugh. I rather dislike tarballs that don't have the package name in them. 
Download it, some time passes, and you look at it and aren't sure what 
it actually is anymore just by looking at the file name. It causes some 
annoyances with rpm specs too (can't use standard %setup macro, because 
it expects %{name}-%{version}, etc).

I'll re-spin the patch series in the morning w/updates from further 
discussion here.

-- 
Jarod Wilson
jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-core 0/6] redhat/spec: various fixes and cleanups

[Doug Ledford]

[-- Attachment #1.1: Type: text/plain, Size: 1139 bytes --]

On 12/21/2016 5:08 PM, Jarod Wilson wrote:
> This is a set of patches to fix up a few things in the Red Hat (Fedora)
> rdma-core spec file that came up as part of the Fedora new package review
> process, where we're working on getting rdma-core into Fedora in place of
> the old split packages.
> 
> Fedora package review bug:
>   https://bugzilla.redhat.com/show_bug.cgi?id=1404043
> 
> Jarod Wilson (6):
>   redhat/spec: call it srp_daemon in desc, not srptools
>   redhat/spec: improve base package description
>   redhat/spec: use fully versioned Requires on base package
>   redhat/spec: fix up %post scriptlets
>   redhat/spec: clean up rsocket packaging
>   redhat/spec: make sure docdir is owned
> 
>  redhat/rdma-core.spec | 65 +++++++++++++++++++++++++++++++++++++--------------
>  1 file changed, 47 insertions(+), 18 deletions(-)
> 

Thanks, series applied and -rc3 generated, so future fixes will need to
be incremental.

-- 
Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
    GPG Key ID: B826A3330E572FDD
    Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 884 bytes --]

Re: [PATCH rdma-core 0/6] redhat/spec: various fixes and cleanups

[Jarod Wilson]

On 2016-12-22 11:53 AM, Doug Ledford wrote:
> On 12/21/2016 5:08 PM, Jarod Wilson wrote:
>> This is a set of patches to fix up a few things in the Red Hat (Fedora)
>> rdma-core spec file that came up as part of the Fedora new package review
>> process, where we're working on getting rdma-core into Fedora in place of
>> the old split packages.
>>
>> Fedora package review bug:
>>   https://bugzilla.redhat.com/show_bug.cgi?id=1404043
>>
>> Jarod Wilson (6):
>>   redhat/spec: call it srp_daemon in desc, not srptools
>>   redhat/spec: improve base package description
>>   redhat/spec: use fully versioned Requires on base package
>>   redhat/spec: fix up %post scriptlets
>>   redhat/spec: clean up rsocket packaging
>>   redhat/spec: make sure docdir is owned
>>
>>  redhat/rdma-core.spec | 65 +++++++++++++++++++++++++++++++++++++--------------
>>  1 file changed, 47 insertions(+), 18 deletions(-)
>>
>
> Thanks, series applied and -rc3 generated, so future fixes will need to
> be incremental.

There are at least two fixes required, but I'd already rebased my local 
tree to include them, merged with patches, so I'll have to extract them. 
Doesn't look like these have been pushed in git just yet? (Ah, 
timestamps says it's only been a few minutes)

-- 
Jarod Wilson
jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-core 5/6] redhat/spec: clean up rsocket packaging

[Jason Gunthorpe]

On Thu, Dec 22, 2016 at 02:17:52AM -0500, Jarod Wilson wrote:
> > %files -n librdmacm
> > %{_libdir}/librdmacm*.so.*
> >-%{_libdir}/rsocket/*.so.*
> >+%dir %{_libdir}/rsocket
> >+%{_libdir}/rsocket/*.so
> 
> Minor error here, should be rsocket/*.so* to pick up the symlinks too, at
> least until they get neutered.

FWIW, it would also be appropriate to put the LDPRELOAD library into
librdmacm-utils..

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-core 0/6] redhat/spec: various fixes and cleanups

[Jason Gunthorpe]

On Thu, Dec 22, 2016 at 02:32:07AM -0500, Jarod Wilson wrote:

> >>BuildRequires: valgrind-devel
> >>honli: systemd and dracut also needed.
> >
> >Not to build?
> 
> Argh. From https://fedoraproject.org/wiki/Packaging:Systemd

Hum, I wonder if that works from docker containers.. C7 was doing
something wonky with fake systemd packages in docker.

> >>honli: Please add "Requires:" entries against rsyslog, systemd, kmod,
> >>honli: logrotate, initscripts, and dracut.
> >
> >It might make sense to depend on kmod in rdma core since it is using
> >the module loading infrastructure..
> 
> The kernel itself already requires it, so unless you're installing on a
> system that somehow has no kernel installed, it's entirely redundant.

Containers don't have the kernel packages.

Containers won't trigger udev rules, but they might be using
systemd.

I also have some Xen systems that boot without kernel packages (shared
kernel comes from the dom0), they will trigger udev related rules and
module loading paths, the scripts shouldn't outright fail because kmod
is accidently not installed.

> >Use this for now:
> >
> >https://github.com/linux-rdma/rdma-core/archive/v12-rc2.tar.gz
> >
> >Best to make sure the release process is going to work :)
>
> Ugh. I rather dislike tarballs that don't have the package name in them.
> Download it, some time passes, and you look at it and aren't sure
> what it

Yes.. Somehow libfabric is doing this:

https://github.com/ofiwg/libfabric/releases

It looks like they produced the .tar.gz by hand and then uploaded it
again as an attached release file?

Should we do this too, Doug?

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-core 0/6] redhat/spec: various fixes and cleanups

[Jarod Wilson]

On 2016-12-22 11:53 AM, Doug Ledford wrote:
> On 12/21/2016 5:08 PM, Jarod Wilson wrote:
>> This is a set of patches to fix up a few things in the Red Hat (Fedora)
>> rdma-core spec file that came up as part of the Fedora new package review
>> process, where we're working on getting rdma-core into Fedora in place of
>> the old split packages.
>>
>> Fedora package review bug:
>>   https://bugzilla.redhat.com/show_bug.cgi?id=1404043
>>
>> Jarod Wilson (6):
>>   redhat/spec: call it srp_daemon in desc, not srptools
>>   redhat/spec: improve base package description
>>   redhat/spec: use fully versioned Requires on base package
>>   redhat/spec: fix up %post scriptlets
>>   redhat/spec: clean up rsocket packaging
>>   redhat/spec: make sure docdir is owned
>>
>>  redhat/rdma-core.spec | 65 +++++++++++++++++++++++++++++++++++++--------------
>>  1 file changed, 47 insertions(+), 18 deletions(-)
>>
>
> Thanks, series applied and -rc3 generated, so future fixes will need to
> be incremental.

Okay, I just sent an incremental patch with a number of other spec 
fixups resulting from discussion here and in the review bug.

-- 
Jarod Wilson
jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-core 5/6] redhat/spec: clean up rsocket packaging

[Doug Ledford]

[-- Attachment #1.1: Type: text/plain, Size: 878 bytes --]

On 12/22/2016 12:33 PM, Jason Gunthorpe wrote:
> On Thu, Dec 22, 2016 at 02:17:52AM -0500, Jarod Wilson wrote:
>>> %files -n librdmacm
>>> %{_libdir}/librdmacm*.so.*
>>> -%{_libdir}/rsocket/*.so.*
>>> +%dir %{_libdir}/rsocket
>>> +%{_libdir}/rsocket/*.so
>>
>> Minor error here, should be rsocket/*.so* to pick up the symlinks too, at
>> least until they get neutered.

I hand fixed this when taking the original series BTW...

> FWIW, it would also be appropriate to put the LDPRELOAD library into
> librdmacm-utils..

I'm pretty sure the Fedora standards would want it in librsocket and not
in a -utils package.  The -utils package should be executables and not
libraries.


-- 
Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
    GPG Key ID: B826A3330E572FDD
    Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 884 bytes --]

Re: [PATCH rdma-core 0/6] redhat/spec: various fixes and cleanups

[Doug Ledford]

[-- Attachment #1.1: Type: text/plain, Size: 704 bytes --]

On 12/22/2016 12:38 PM, Jason Gunthorpe wrote:
> On Thu, Dec 22, 2016 at 02:32:07AM -0500, Jarod Wilson wrote:

> Yes.. Somehow libfabric is doing this:
> 
> https://github.com/ofiwg/libfabric/releases
> 
> It looks like they produced the .tar.gz by hand and then uploaded it
> again as an attached release file?
> 
> Should we do this too, Doug?

I was going to try changing the tag format on the next tag to get the
desired result.  So, instead of v12-rc4 or v12, rdma-core-12-rc4 or just
rdma-core-12.


-- 
Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
    GPG Key ID: B826A3330E572FDD
    Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 884 bytes --]

Re: [PATCH rdma-core 5/6] redhat/spec: clean up rsocket packaging

[Jason Gunthorpe]

On Thu, Dec 22, 2016 at 01:15:55PM -0500, Doug Ledford wrote:
> On 12/22/2016 12:33 PM, Jason Gunthorpe wrote:
> > On Thu, Dec 22, 2016 at 02:17:52AM -0500, Jarod Wilson wrote:
> >>> %files -n librdmacm
> >>> %{_libdir}/librdmacm*.so.*
> >>> -%{_libdir}/rsocket/*.so.*
> >>> +%dir %{_libdir}/rsocket
> >>> +%{_libdir}/rsocket/*.so
> >>
> >> Minor error here, should be rsocket/*.so* to pick up the symlinks too, at
> >> least until they get neutered.
> 
> I hand fixed this when taking the original series BTW...
> 
> > FWIW, it would also be appropriate to put the LDPRELOAD library into
> > librdmacm-utils..
> 
> I'm pretty sure the Fedora standards would want it in librsocket and not
> in a -utils package.  The -utils package should be executables and not
> libraries.

Not sure, don't see anything about preload libraries. This is not a
link library, it is only executable via LD_PRELOAD, and it shouldn't
result in any automatic RPM Provides either (ditto for all our
plugins, I didn't check this).

Ideally there would be a 'run under rsockets' helper tool that did the
preload (ala fakeroot), and then it would be clear that the library is
a private data file of the helper and not a system link library.

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-core 5/6] redhat/spec: clean up rsocket packaging

[Jarod Wilson]

On 2016-12-22 1:15 PM, Doug Ledford wrote:
> On 12/22/2016 12:33 PM, Jason Gunthorpe wrote:
>> On Thu, Dec 22, 2016 at 02:17:52AM -0500, Jarod Wilson wrote:
>>>> %files -n librdmacm
>>>> %{_libdir}/librdmacm*.so.*
>>>> -%{_libdir}/rsocket/*.so.*
>>>> +%dir %{_libdir}/rsocket
>>>> +%{_libdir}/rsocket/*.so
>>>
>>> Minor error here, should be rsocket/*.so* to pick up the symlinks too, at
>>> least until they get neutered.
>
> I hand fixed this when taking the original series BTW...

Okay, this fixup was in my follow-up patch as well. Guessing what to 
patch against a little, since it's not pushed yet. ;)

>> FWIW, it would also be appropriate to put the LDPRELOAD library into
>> librdmacm-utils..
>
> I'm pretty sure the Fedora standards would want it in librsocket and not
> in a -utils package.  The -utils package should be executables and not
> libraries.

Yeah, I think it's probably best where it is now.

Do you want a re-respin of the follow-up patch to drop the chmod bit and 
duplicitous librspreload fixup, or can you just smash it all together?

-- 
Jarod Wilson
jarod-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-core 5/6] redhat/spec: clean up rsocket packaging

[Doug Ledford]

[-- Attachment #1.1: Type: text/plain, Size: 1160 bytes --]

On 12/22/2016 1:23 PM, Jarod Wilson wrote:
> On 2016-12-22 1:15 PM, Doug Ledford wrote:
>> On 12/22/2016 12:33 PM, Jason Gunthorpe wrote:
>>> On Thu, Dec 22, 2016 at 02:17:52AM -0500, Jarod Wilson wrote:
>>>>> %files -n librdmacm
>>>>> %{_libdir}/librdmacm*.so.*
>>>>> -%{_libdir}/rsocket/*.so.*
>>>>> +%dir %{_libdir}/rsocket
>>>>> +%{_libdir}/rsocket/*.so
>>>>
>>>> Minor error here, should be rsocket/*.so* to pick up the symlinks
>>>> too, at
>>>> least until they get neutered.
>>
>> I hand fixed this when taking the original series BTW...
> 
> Okay, this fixup was in my follow-up patch as well. Guessing what to
> patch against a little, since it's not pushed yet. ;)

So, that was my mistake.  I pushed using --tags and that sent the tag,
but not the master ref update.  I just pushed out my current master.  I
thought adding --tags to the push command would add tags to the push,
not replace the normal push with only a tag push.  My mistake.


-- 
Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
    GPG Key ID: B826A3330E572FDD
    Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 884 bytes --]

Re: [PATCH rdma-core 0/6] redhat/spec: various fixes and cleanups

[Jason Gunthorpe]

On Thu, Dec 22, 2016 at 01:17:52PM -0500, Doug Ledford wrote:
> On 12/22/2016 12:38 PM, Jason Gunthorpe wrote:
> > On Thu, Dec 22, 2016 at 02:32:07AM -0500, Jarod Wilson wrote:
> 
> > Yes.. Somehow libfabric is doing this:
> > 
> > https://github.com/ofiwg/libfabric/releases
> > 
> > It looks like they produced the .tar.gz by hand and then uploaded it
> > again as an attached release file?
> > 
> > Should we do this too, Doug?
> 
> I was going to try changing the tag format on the next tag to get the
> desired result.  So, instead of v12-rc4 or v12, rdma-core-12-rc4 or just
> rdma-core-12.

I don't know enough about github to comment..

The advantage of the libfabric approach is that the tar.gz is produced
once and never changed so we can provide the SHA2, while I'm not
certain the github .tar.gz link is bit-for-bit content-stable?

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-core 0/6] redhat/spec: various fixes and cleanups

[Doug Ledford]

[-- Attachment #1.1: Type: text/plain, Size: 1371 bytes --]

On 12/22/2016 3:28 PM, Jason Gunthorpe wrote:
> On Thu, Dec 22, 2016 at 01:17:52PM -0500, Doug Ledford wrote:
>> On 12/22/2016 12:38 PM, Jason Gunthorpe wrote:
>>> On Thu, Dec 22, 2016 at 02:32:07AM -0500, Jarod Wilson wrote:
>>
>>> Yes.. Somehow libfabric is doing this:
>>>
>>> https://github.com/ofiwg/libfabric/releases
>>>
>>> It looks like they produced the .tar.gz by hand and then uploaded it
>>> again as an attached release file?
>>>
>>> Should we do this too, Doug?
>>
>> I was going to try changing the tag format on the next tag to get the
>> desired result.  So, instead of v12-rc4 or v12, rdma-core-12-rc4 or just
>> rdma-core-12.
> 
> I don't know enough about github to comment..
> 
> The advantage of the libfabric approach is that the tar.gz is produced
> once and never changed so we can provide the SHA2, while I'm not
> certain the github .tar.gz link is bit-for-bit content-stable?

Yeah, but the SHA hash of the git repo is, for all intents and purposes,
just as valid as a SHA for the tarball.  It's just verified in a
different way.  I can certainly see the benefit of the libfabric
approach for keeping the status quo happy though.


-- 
Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
    GPG Key ID: B826A3330E572FDD
    Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 884 bytes --]