diff mbox

[dpdk-dev,v2] crypto/dpaa2_sec: fix the return of supported API

Message ID 20170718073356.14288-1-akhil.goyal@nxp.com (mailing list archive)
State Not Applicable, archived
Delegated to: Pablo de Lara Guarch
Headers

Checks

Context Check Description
ci/checkpatch success coding style OK
ci/Intel-compilation success Compilation OK

Commit Message

Akhil Goyal July 18, 2017, 7:33 a.m. UTC 
  call to dpaa2_sec_dev_configure() is made mandatory,
but dpaa2_sec_pmd returns a ENOTSUP which results in
device not getting configured.

dpaa2_sec PMD does not need any further configuration
to be done in dpaa2_sec_dev_configure, hence returning 0

Fixes: e5cbdfc53765 ("crypto/dpaa2_sec: add basic operations")

Cc: stable@dpdk.org

Signed-off-by: Akhil Goyal <akhil.goyal@nxp.com>
---
This patch is for stable 17.05 tree only.
This fix is already present in dpdk mainline

 drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

De Lara Guarch, Pablo July 18, 2017, 10:26 a.m. UTC | #1 
Hi Akhil,

> -----Original Message-----
> From: Akhil Goyal [mailto:akhil.goyal@nxp.com]
> Sent: Tuesday, July 18, 2017 8:34 AM
> To: dev@dpdk.org; stable@dpdk.org
> Cc: hemant.agrawal@nxp.com; De Lara Guarch, Pablo
> <pablo.de.lara.guarch@intel.com>; yliu@fridaylinux.org; Akhil Goyal
> <akhil.goyal@nxp.com>
> Subject: [PATCH v2] crypto/dpaa2_sec: fix the return of supported API
> 
> call to dpaa2_sec_dev_configure() is made mandatory, but dpaa2_sec_pmd
> returns a ENOTSUP which results in device not getting configured.
> 
> dpaa2_sec PMD does not need any further configuration to be done in
> dpaa2_sec_dev_configure, hence returning 0
> 
> Fixes: e5cbdfc53765 ("crypto/dpaa2_sec: add basic operations")
> 
> Cc: stable@dpdk.org
> 
> Signed-off-by: Akhil Goyal <akhil.goyal@nxp.com>

Looks ok to me, but this is only applicable in the stable branch,
so no need to send it to dev@dpdk.org.

Thanks,
Pablo
Yuanhan Liu July 18, 2017, 2:39 p.m. UTC | #2 
On Tue, Jul 18, 2017 at 10:26:39AM +0000, De Lara Guarch, Pablo wrote:
> Hi Akhil,
> 
> > -----Original Message-----
> > From: Akhil Goyal [mailto:akhil.goyal@nxp.com]
> > Sent: Tuesday, July 18, 2017 8:34 AM
> > To: dev@dpdk.org; stable@dpdk.org
> > Cc: hemant.agrawal@nxp.com; De Lara Guarch, Pablo
> > <pablo.de.lara.guarch@intel.com>; yliu@fridaylinux.org; Akhil Goyal
> > <akhil.goyal@nxp.com>
> > Subject: [PATCH v2] crypto/dpaa2_sec: fix the return of supported API
> > 
> > call to dpaa2_sec_dev_configure() is made mandatory, but dpaa2_sec_pmd
> > returns a ENOTSUP which results in device not getting configured.
> > 
> > dpaa2_sec PMD does not need any further configuration to be done in
> > dpaa2_sec_dev_configure, hence returning 0
> > 
> > Fixes: e5cbdfc53765 ("crypto/dpaa2_sec: add basic operations")
> > 
> > Cc: stable@dpdk.org
> > 
> > Signed-off-by: Akhil Goyal <akhil.goyal@nxp.com>
> 
> Looks ok to me, but this is only applicable in the stable branch,
> so no need to send it to dev@dpdk.org.

Why? We already have such fix in upstream? Normally, we just pick upstream
commits (but not patches: the emails) to stable release.

	--yliu
De Lara Guarch, Pablo July 18, 2017, 4:21 p.m. UTC | #3 
HI Yuanhan,

> -----Original Message-----
> From: Yuanhan Liu [mailto:yliu@fridaylinux.org]
> Sent: Tuesday, July 18, 2017 3:39 PM
> To: De Lara Guarch, Pablo <pablo.de.lara.guarch@intel.com>
> Cc: Akhil Goyal <akhil.goyal@nxp.com>; dev@dpdk.org; stable@dpdk.org;
> hemant.agrawal@nxp.com
> Subject: Re: [PATCH v2] crypto/dpaa2_sec: fix the return of supported API
> 
> On Tue, Jul 18, 2017 at 10:26:39AM +0000, De Lara Guarch, Pablo wrote:
> > Hi Akhil,
> >
> > > -----Original Message-----
> > > From: Akhil Goyal [mailto:akhil.goyal@nxp.com]
> > > Sent: Tuesday, July 18, 2017 8:34 AM
> > > To: dev@dpdk.org; stable@dpdk.org
> > > Cc: hemant.agrawal@nxp.com; De Lara Guarch, Pablo
> > > <pablo.de.lara.guarch@intel.com>; yliu@fridaylinux.org; Akhil Goyal
> > > <akhil.goyal@nxp.com>
> > > Subject: [PATCH v2] crypto/dpaa2_sec: fix the return of supported
> > > API
> > >
> > > call to dpaa2_sec_dev_configure() is made mandatory, but
> > > dpaa2_sec_pmd returns a ENOTSUP which results in device not getting
> configured.
> > >
> > > dpaa2_sec PMD does not need any further configuration to be done in
> > > dpaa2_sec_dev_configure, hence returning 0
> > >
> > > Fixes: e5cbdfc53765 ("crypto/dpaa2_sec: add basic operations")
> > >
> > > Cc: stable@dpdk.org
> > >
> > > Signed-off-by: Akhil Goyal <akhil.goyal@nxp.com>
> >
> > Looks ok to me, but this is only applicable in the stable branch, so
> > no need to send it to dev@dpdk.org.
> 
> Why? We already have such fix in upstream? Normally, we just pick
> upstream commits (but not patches: the emails) to stable release.

It looks like this fix was included in
13273250eec5 ("crypto/dpaa2_sec: support AES-GCM and CTR").
Unfortunately, this patch should have been split into two different patches.
Since this has already been merged, I think our only way to integrate this
In 17.05.1 is by getting it separately.

Pablo

> 
> 	--yliu
Hemant Agrawal July 18, 2017, 5:32 p.m. UTC | #4 
HI Yuanhan,

> > > > -----Original Message-----
> > > > From: Akhil Goyal [mailto:akhil.goyal@nxp.com]
> > > > Sent: Tuesday, July 18, 2017 8:34 AM
> > > > To: dev@dpdk.org; stable@dpdk.org
> > > > Cc: hemant.agrawal@nxp.com; De Lara Guarch, Pablo
> > > > <pablo.de.lara.guarch@intel.com>; yliu@fridaylinux.org; Akhil
> > > > Goyal <akhil.goyal@n[Hemant] axp.com>
> > > > Subject: [PATCH v2] crypto/dpaa2_sec: fix the return of supported
> > > > API
> > > >
> > > > call to dpaa2_sec_dev_configure() is made mandatory, but
> > > > dpaa2_sec_pmd returns a ENOTSUP which results in device not
> > > > getting
> > configured.
> > > >
> > > > dpaa2_sec PMD does not need any further configuration to be done
> > > > in dpaa2_sec_dev_configure, hence returning 0
> > > >
> > > > Fixes: e5cbdfc53765 ("crypto/dpaa2_sec: add basic operations")
> > > >
> > > > Cc: stable@dpdk.org
> > > >
> > > > Signed-off-by: Akhil Goyal <akhil.goyal@nxp.com>
> > >
> > > Looks ok to me, but this is only applicable in the stable branch, so
> > > no need to send it to dev@dpdk.org.
> >
> > Why? We already have such fix in upstream? Normally, we just pick
> > upstream commits (but not patches: the emails) to stable release.
> 
> It looks like this fix was included in
> 13273250eec5 ("crypto/dpaa2_sec: support AES-GCM and CTR").
> Unfortunately, this patch should have been split into two different patches.
> Since this has already been merged, I think our only way to integrate this In
> 17.05.1 is by getting it separately.

In general, there may be other incidents, where a patch is only applicable for the stable tree. It may not be applicable for upstream tree due to architecture changes or other reasons.
How do you want to handle such patches? 

e.g. in OVS, we can do it by marking the patch header with "[branch-2.6]"

Regards,
Hemant
Yuanhan Liu July 19, 2017, 12:12 p.m. UTC | #5 
On Tue, Jul 18, 2017 at 05:32:44PM +0000, Hemant Agrawal wrote:
> > > > > Subject: [PATCH v2] crypto/dpaa2_sec: fix the return of supported
> > > > > API
> > > > >
> > > > > call to dpaa2_sec_dev_configure() is made mandatory, but
> > > > > dpaa2_sec_pmd returns a ENOTSUP which results in device not
> > > > > getting
> > > configured.
> > > > >
> > > > > dpaa2_sec PMD does not need any further configuration to be done
> > > > > in dpaa2_sec_dev_configure, hence returning 0
> > > > >
> > > > > Fixes: e5cbdfc53765 ("crypto/dpaa2_sec: add basic operations")
> > > > >
> > > > > Cc: stable@dpdk.org
> > > > >
> > > > > Signed-off-by: Akhil Goyal <akhil.goyal@nxp.com>
> > > >
> > > > Looks ok to me, but this is only applicable in the stable branch, so
> > > > no need to send it to dev@dpdk.org.
> > >
> > > Why? We already have such fix in upstream? Normally, we just pick
> > > upstream commits (but not patches: the emails) to stable release.
> > 
> > It looks like this fix was included in
> > 13273250eec5 ("crypto/dpaa2_sec: support AES-GCM and CTR").
> > Unfortunately, this patch should have been split into two different patches.
> > Since this has already been merged, I think our only way to integrate this In
> > 17.05.1 is by getting it separately.
> 
> In general, there may be other incidents, where a patch is only applicable for the stable tree. It may not be applicable for upstream tree due to architecture changes or other reasons.
> How do you want to handle such patches? 
> 
> e.g. in OVS, we can do it by marking the patch header with "[branch-2.6]"


Yes, you are right, it might happen. Then you need cook a standalone
patch and send it to stable ml only. Since I don't usually pick stable
patches directly from stable ML, you probably need add some marks in
the commit log. Something like "this is for stable tree only and add
a bit explanation".

Normally, every time I saw a patch sent only to stable ML I will ask
the same question I have asked in this email. But I could just miss
it. So you are suggested to do above.

For this case, just as Pablo said, the patch should be split in the
beginning, then only the (small) bug fixing patch will be picked to a
specific stable release. And since it already happened, you could just
send it to stable ML only, and better, with me cc-ed.

Thanks.

	--yliu
diff mbox

Patch

diff --git a/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c b/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
index 4e01fe8..08fd085 100644
--- a/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
+++ b/drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c
@@ -1263,7 +1263,7 @@  dpaa2_sec_dev_configure(struct rte_cryptodev *dev __rte_unused,
 {
 	PMD_INIT_FUNC_TRACE();
 
-	return -ENOTSUP;
+	return 0;
 }
 
 static int