[dpdk-dev] [EXT] Re: [PATCH] devtools: skip the symbol check when map file under drivers

Neil Horman nhorman at tuxdriver.com
Wed May 22 14:45:33 CEST 2019


On Wed, May 22, 2019 at 03:05:54AM +0000, Jerin Jacob Kollanukkaran wrote:
> > -----Original Message-----
> > From: Neil Horman <nhorman at tuxdriver.com>
> > Sent: Wednesday, May 22, 2019 1:57 AM
> > To: Jerin Jacob Kollanukkaran <jerinj at marvell.com>
> > Cc: dev at dpdk.org; thomas at monjalon.net; stable at dpdk.org
> > Subject: [EXT] Re: [dpdk-dev] [PATCH] devtools: skip the symbol check when
> > map file under drivers
> > 
> > On Wed, May 22, 2019 at 01:26:28AM +0530, jerinj at marvell.com wrote:
> > > From: Jerin Jacob <jerinj at marvell.com>
> > >
> > > Drivers do not have ABI.
> > > Skip the symbol check if map file under drivers directory.
> > >
> > > Fixes: 4bec48184e33 ("devtools: add checks for ABI symbol addition")
> > >
> > > Cc: stable at dpdk.org
> > > Cc: Neil Horman <nhorman at tuxdriver.com>
> > >
> > Sorry, but I'm not ok with this, because many of our DPDK PMDs have functions
> > that get exported which are meant to be called by applications directly.  The
> 
> OK. Just to update my knowledge, Should those API needs to go through ABI/API
> depreciation process?
> 
Yes, they definately should, they are API's just as any other in the core DPDK
library.

> Actually, I am concerned about the APIs, which is called between drviers not
> the application. For example,
> drivers/common/dpaax/rte_common_dpaax_version.map
> 
> it is not interface to application rather it is for intra driver case.
> I think, I can change my logic to Skip the symbols which NOT starting with rte_.
> Agree?
> 
No, Thats just one case, and if those calls are between drivers, so be it, but
those still need to be stable, and we have other examples (like the bonding or
dummy driver), which have additional APIs that are explicitly meant to be used
by an application.

> Context:
> I am adding a new driver/common/octeontx2 directory and it has some API which
> Needs to shared between drivers not to the application. For me, it does not make
> sense to go through any ABI process in such case.
> 
Why?  If you create an API thats reachable from another block of code (be it a
driver or an application), you've created a dependency in which that API must
remain stable, lest you risk breaking something.  If an end user writes an
out-of-tree PMD which makes use of an an additional driver API, then you need to
keep it stable or you will break them.

> 
> > dpaa2 driver is a good example, the cryptodev scheduler is another.  Take a look
> > at their version.map files to see what I mean.
> > 
> > Unless we are willing to make drivers opaque objects that are only accessible
> > from the [eth|crypto|etc]dev apis in the DPDK core, we have the potential for
> > exported symbols, which means we have an ABI that has to be maintained, or at
> > least recognized and reviewed for consistency
> > 
> > Nacked-by: Neil Horman <nhorman at tuxdriver.com>
> > 
> > 
> > > Signed-off-by: Jerin Jacob <jerinj at marvell.com>
> > > ---
> > >  devtools/check-symbol-change.sh | 8 ++++++++
> > >  1 file changed, 8 insertions(+)
> > >
> > > diff --git a/devtools/check-symbol-change.sh
> > > b/devtools/check-symbol-change.sh index c5434f3bb..444beddad 100755
> > > --- a/devtools/check-symbol-change.sh
> > > +++ b/devtools/check-symbol-change.sh
> > > @@ -93,6 +93,14 @@ check_for_rule_violations()
> > >  		if [ "$ar" = "add" ]
> > >  		then
> > >
> > > +			directory=`echo $mname | cut -d / -f 2`
> > > +			if [ "$directory" = "drivers" ]
> > > +			then
> > > +				# Drivers do not have ABI. Skip further
> > > +				# processing if the map file is under
> > > +				# drivers directory
> > > +				continue
> > > +			fi
> > >  			if [ "$secname" = "unknown" ]
> > >  			then
> > >  				# Just inform the user of this occurrence, but
> > > --
> > > 2.21.0
> > >
> > >
> 


More information about the dev mailing list