[dpdk-dev,v2] mk: Provide option to set Major ABI version

  Downstreams might want to provide different DPDK releases at the same
time to support multiple consumers of DPDK linked against older and newer
sonames.

Also due to the interdependencies that DPDK libraries can have applications
might end up with an executable space in which multiple versions of a
library are mapped by ld.so.

Think of LibA that got an ABI bump and LibB that did not get an ABI bump
but is depending on LibA.

    Application
    \-> LibA.old
    \-> LibB.new -> LibA.new

That is a conflict which can be avoided by setting CONFIG_RTE_MAJOR_ABI.
If set CONFIG_RTE_MAJOR_ABI overwrites any LIBABIVER value.
An example might be ``CONFIG_RTE_MAJOR_ABI=16.11`` which will make all
libraries librte<?>.so.16.11 instead of librte<?>.so.<LIBABIVER>.

We need to cut arbitrary long stings after the .so now and this would work
for any ABI version in LIBABIVER:
  $(Q)ln -s -f $< $(patsubst %.$(LIBABIVER),%,$@)
But using the following instead additionally allows to simplify the Make
File for the CONFIG_RTE_NEXT_ABI case.
  $(Q)ln -s -f $< $(shell echo $@ | sed 's/\.so.*/.so/')

Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
---
 config/common_base                     |  5 +++++
 doc/guides/contributing/versioning.rst | 25 +++++++++++++++++++++++++
 mk/rte.lib.mk                          | 14 +++++++++-----
 3 files changed, 39 insertions(+), 5 deletions(-)
  

On Wed, Mar 1, 2017 at 10:34 AM, Christian Ehrhardt
<christian.ehrhardt@canonical.com> wrote:
> Downstreams might want to provide different DPDK releases at the same
> time to support multiple consumers of DPDK linked against older and newer
> sonames.
>
> Also due to the interdependencies that DPDK libraries can have applications
> might end up with an executable space in which multiple versions of a
> library are mapped by ld.so.
>
> Think of LibA that got an ABI bump and LibB that did not get an ABI bump
> but is depending on LibA.
>
>     Application
>     \-> LibA.old
>     \-> LibB.new -> LibA.new
>
> That is a conflict which can be avoided by setting CONFIG_RTE_MAJOR_ABI.
> If set CONFIG_RTE_MAJOR_ABI overwrites any LIBABIVER value.
> An example might be ``CONFIG_RTE_MAJOR_ABI=16.11`` which will make all
> libraries librte<?>.so.16.11 instead of librte<?>.so.<LIBABIVER>.
>
> We need to cut arbitrary long stings after the .so now and this would work
> for any ABI version in LIBABIVER:
>   $(Q)ln -s -f $< $(patsubst %.$(LIBABIVER),%,$@)
> But using the following instead additionally allows to simplify the Make
> File for the CONFIG_RTE_NEXT_ABI case.
>   $(Q)ln -s -f $< $(shell echo $@ | sed 's/\.so.*/.so/')
>
> Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
> ---
>  config/common_base                     |  5 +++++
>  doc/guides/contributing/versioning.rst | 25 +++++++++++++++++++++++++
>  mk/rte.lib.mk                          | 14 +++++++++-----
>  3 files changed, 39 insertions(+), 5 deletions(-)
>
> diff --git a/config/common_base b/config/common_base
> index aeee13e..37aa1e1 100644
> --- a/config/common_base
> +++ b/config/common_base
> @@ -75,6 +75,11 @@ CONFIG_RTE_BUILD_SHARED_LIB=n
>  CONFIG_RTE_NEXT_ABI=y
>
>  #
> +# Major ABI to overwrite library specific LIBABIVER
> +#
> +CONFIG_RTE_MAJOR_ABI=
> +
> +#
>  # Machine's cache line size
>  #
>  CONFIG_RTE_CACHE_LINE_SIZE=64
> diff --git a/doc/guides/contributing/versioning.rst b/doc/guides/contributing/versioning.rst
> index fbc44a7..8aaf370 100644
> --- a/doc/guides/contributing/versioning.rst
> +++ b/doc/guides/contributing/versioning.rst
> @@ -133,6 +133,31 @@ The macros exported are:
>    fully qualified function ``p``, so that if a symbol becomes versioned, it
>    can still be mapped back to the public symbol name.
>
> +Setting a Major ABI version
> +---------------------------
> +
> +Downstreams might want to provide different DPDK releases at the same time to
> +support multiple consumers of DPDK linked against older and newer sonames.
> +
> +Also due to the interdependencies that DPDK libraries can have applications
> +might end up with an executable space in which multiple versions of a library
> +are mapped by ld.so.
> +
> +Think of LibA that got an ABI bump and LibB that did not get an ABI bump but is
> +depending on LibA.
> +
> +.. note::
> +
> +    Application
> +    \-> LibA.old
> +    \-> LibB.new -> LibA.new
> +
> +That is a conflict which can be avoided by setting ``CONFIG_RTE_MAJOR_ABI``.
> +If set, the value of ``CONFIG_RTE_MAJOR_ABI`` overwrites all - otherwise per
> +library - versions defined in the libraries ``LIBABIVER``.
> +An example might be ``CONFIG_RTE_MAJOR_ABI=16.11`` which will make all libraries
> +``librte<?>.so.16.11`` instead of ``librte<?>.so.<LIBABIVER>``.
> +
>  Examples of ABI Macro use
>  -------------------------
>
> diff --git a/mk/rte.lib.mk b/mk/rte.lib.mk
> index 33a5f5a..1ffbf42 100644
> --- a/mk/rte.lib.mk
> +++ b/mk/rte.lib.mk
> @@ -40,12 +40,20 @@ EXTLIB_BUILD ?= n
>  # VPATH contains at least SRCDIR
>  VPATH += $(SRCDIR)
>
> +ifneq ($(CONFIG_RTE_MAJOR_ABI),)
> +ifneq ($(LIBABIVER),)
> +LIBABIVER := $(CONFIG_RTE_MAJOR_ABI)
> +endif
> +endif
> +
>  ifeq ($(CONFIG_RTE_BUILD_SHARED_LIB),y)
>  LIB := $(patsubst %.a,%.so.$(LIBABIVER),$(LIB))
>  ifeq ($(EXTLIB_BUILD),n)
> +ifeq ($(CONFIG_RTE_MAJOR_ABI),)
>  ifeq ($(CONFIG_RTE_NEXT_ABI),y)
>  LIB := $(LIB).1
>  endif
> +endif
>  CPU_LDFLAGS += --version-script=$(SRCDIR)/$(EXPORT_MAP)
>  endif
>  endif
> @@ -156,11 +164,7 @@ $(RTE_OUTPUT)/lib/$(LIB): $(LIB)
>         @[ -d $(RTE_OUTPUT)/lib ] || mkdir -p $(RTE_OUTPUT)/lib
>         $(Q)cp -f $(LIB) $(RTE_OUTPUT)/lib
>  ifeq ($(CONFIG_RTE_BUILD_SHARED_LIB),y)
> -ifeq ($(CONFIG_RTE_NEXT_ABI)$(EXTLIB_BUILD),yn)
> -       $(Q)ln -s -f $< $(basename $(basename $@))
> -else
> -       $(Q)ln -s -f $< $(basename $@)
> -endif
> +       $(Q)ln -s -f $< $(shell echo $@ | sed 's/\.so.*/.so/')
>  endif
>
>  #
> --
> 2.7.4
>

Reviewed-by: Jan Blunck <jblunck@infradead.org>
Tested-by: Jan Blunck <jblunck@infradead.org>
  

2017-03-01 15:35, Jan Blunck:
> On Wed, Mar 1, 2017 at 10:34 AM, Christian Ehrhardt
> <christian.ehrhardt@canonical.com> wrote:
> > Downstreams might want to provide different DPDK releases at the same
> > time to support multiple consumers of DPDK linked against older and newer
> > sonames.
> >
> > Also due to the interdependencies that DPDK libraries can have applications
> > might end up with an executable space in which multiple versions of a
> > library are mapped by ld.so.
> >
> > Think of LibA that got an ABI bump and LibB that did not get an ABI bump
> > but is depending on LibA.
> >
> >     Application
> >     \-> LibA.old
> >     \-> LibB.new -> LibA.new
> >
> > That is a conflict which can be avoided by setting CONFIG_RTE_MAJOR_ABI.
> > If set CONFIG_RTE_MAJOR_ABI overwrites any LIBABIVER value.
> > An example might be ``CONFIG_RTE_MAJOR_ABI=16.11`` which will make all
> > libraries librte<?>.so.16.11 instead of librte<?>.so.<LIBABIVER>.
[...]
> >
> > Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
> 
> Reviewed-by: Jan Blunck <jblunck@infradead.org>
> Tested-by: Jan Blunck <jblunck@infradead.org>

Not sure about how it can be used in distributions, but it does not hurt
to provide the config option.
Are you going to link applications against a fixed DPDK version for
every libraries?

Applied, thanks
  

On Thu, Mar 16, 2017 at 6:19 PM, Thomas Monjalon <thomas.monjalon@6wind.com>
wrote:

> Not sure about how it can be used in distributions, but it does not hurt
> to provide the config option.
> Are you going to link applications against a fixed DPDK version for
> every libraries?
>

Kind of yes - we can still update "inside" a major version e.g. stable
releases just fine.
That helps a lot transitioning from one to the next release.

I have an RFC patch for the Debian packaging making use of it this function
that we will likely refresh and pick up on our next merge of a DPDK version.
And I already heard that even other downstreams are using it to simplify
their dependencies.

Applied, thanks
>

Thanks !
  

On Fri, Mar 17, 2017 at 9:27 AM, Christian Ehrhardt
<christian.ehrhardt@canonical.com> wrote:
>
> On Thu, Mar 16, 2017 at 6:19 PM, Thomas Monjalon <thomas.monjalon@6wind.com>
> wrote:
>>
>> Not sure about how it can be used in distributions, but it does not hurt
>> to provide the config option.
>> Are you going to link applications against a fixed DPDK version for
>> every libraries?
>
>
> Kind of yes - we can still update "inside" a major version e.g. stable
> releases just fine.
> That helps a lot transitioning from one to the next release.

Exactly. It enables me to roll out a new major release without being
blocked on the (external) consumers to pick up the changes. Now I can
do this one by one with reduced risk because the runtimes are clearly
separated.

> And I already heard that even other downstreams are using it to simplify
> their dependencies.
>

Downstream?! Pha! ;)