Message ID | 20250408030807.1978102-1-denis@denix.org |
---|---|
State | Accepted |
Delegated to: | Ryan Eatmon |
Headers | show |
Series | [master] optee-client: remove patch that got accepted upstream | expand |
I was unsure about doing this... meta-arm added a second optee recipe version. So both exist: optee-client_4.3.0.bb optee-client_4.4.0.bb So "theoretically" the build should prefer the newer version, right? But for some reason, our builds keep loading the 4.3.0 recipe and then our bbappend bumps the version to 4.4.0. The 4.3.0 recipe has this offending patch, but the 4.4.0 version does not. I've been trying to figure out why our builds keep locking into 4.3.0 when the 4.4.0 is sitting right there... I agree that this patch is a good work around, but I'm not 100% sure it's the final best fix. On 4/7/2025 10:08 PM, Denys Dmytriyenko wrote: > From: Denys Dmytriyenko <denys@konsulko.com> > > meta-arm adds a patch that was sent and accepted upstream, remove > it since mta-ti pulls a more recent verstion. > > Signed-off-by: Denys Dmytriyenko <denys@konsulko.com> > --- > .../recipes-security/optee/optee-client-ti-version.inc | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc b/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc > index 2c8977dd..a628e446 100644 > --- a/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc > +++ b/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc > @@ -1,2 +1,6 @@ > PV = "4.4.0+git" > SRCREV = "d221676a58b305bddbf97db00395205b3038de8e" > + > +SRC_URI:remove = " \ > + file://0001-tee-supplicant-add-udev-rule-and-systemd-service-fil.patch \ > +" > > > > -=-=-=-=-=-=-=-=-=-=-=- > Links: You receive all messages sent to this group. > View/Reply Online (#18447): https://lists.yoctoproject.org/g/meta-ti/message/18447 > Mute This Topic: https://lists.yoctoproject.org/mt/112146993/6551054 > Group Owner: meta-ti+owner@lists.yoctoproject.org > Unsubscribe: https://lists.yoctoproject.org/g/meta-ti/unsub [reatmon@ti.com] > -=-=-=-=-=-=-=-=-=-=-=- >
On 4/8/2025 8:41 AM, Ryan Eatmon via lists.yoctoproject.org wrote: > > I was unsure about doing this... > > meta-arm added a second optee recipe version. So both exist: > optee-client_4.3.0.bb > optee-client_4.4.0.bb > > So "theoretically" the build should prefer the newer version, right? > > But for some reason, our builds keep loading the 4.3.0 recipe and then > our bbappend bumps the version to 4.4.0. The 4.3.0 recipe has this > offending patch, but the 4.4.0 version does not. > > I've been trying to figure out why our builds keep locking into 4.3.0 > when the 4.4.0 is sitting right there... > > I agree that this patch is a good work around, but I'm not 100% sure > it's the final best fix. > It looks like our setting of PV in our bbappend is what is causing this. When I comment out that line the correct recipe version from meta-arm is picked. But when I change it from "4.4.0+git" to just "4.4.0" it still picks the 4.3.0 version. It's like setting that make bitbake think that the 4.3.0 recipe is newer (or as new) as the 4.4.0 recipe. Setting PV = "4.3.9" also does not work. Looks like setting PV to anything in our .inc file causes this bad behavior. Thoughts? > > > On 4/7/2025 10:08 PM, Denys Dmytriyenko wrote: >> From: Denys Dmytriyenko <denys@konsulko.com> >> >> meta-arm adds a patch that was sent and accepted upstream, remove >> it since mta-ti pulls a more recent verstion. >> >> Signed-off-by: Denys Dmytriyenko <denys@konsulko.com> >> --- >> .../recipes-security/optee/optee-client-ti-version.inc | 4 ++++ >> 1 file changed, 4 insertions(+) >> >> diff --git >> a/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc >> b/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc >> index 2c8977dd..a628e446 100644 >> --- a/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc >> +++ b/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc >> @@ -1,2 +1,6 @@ >> PV = "4.4.0+git" >> SRCREV = "d221676a58b305bddbf97db00395205b3038de8e" >> + >> +SRC_URI:remove = " \ >> + >> file://0001-tee-supplicant-add-udev-rule-and-systemd-service-fil.patch \ >> +" >> >> >> >> >> > > > > -=-=-=-=-=-=-=-=-=-=-=- > Links: You receive all messages sent to this group. > View/Reply Online (#18449): https://lists.yoctoproject.org/g/meta-ti/message/18449 > Mute This Topic: https://lists.yoctoproject.org/mt/112146993/6551054 > Group Owner: meta-ti+owner@lists.yoctoproject.org > Unsubscribe: https://lists.yoctoproject.org/g/meta-ti/unsub [reatmon@ti.com] > -=-=-=-=-=-=-=-=-=-=-=- >
On Tue, Apr 08, 2025 at 08:41:20AM -0500, Ryan Eatmon wrote: > > I was unsure about doing this... > > meta-arm added a second optee recipe version. So both exist: > optee-client_4.3.0.bb > optee-client_4.4.0.bb > > So "theoretically" the build should prefer the newer version, right? No, that's not how it works. Our bbappend in version-agnostic and hence gets applied to both base recipes. And then bbappend bumps the PV and they are now matching and neither is newer than the other. > But for some reason, our builds keep loading the 4.3.0 recipe and > then our bbappend bumps the version to 4.4.0. The 4.3.0 recipe has > this offending patch, but the 4.4.0 version does not. > > I've been trying to figure out why our builds keep locking into > 4.3.0 when the 4.4.0 is sitting right there... It is a downside of version-agnostic bbappends with PV changes - we've been in this situation before. And we used some version-specific tricks to make it work. Another solution is to make bbappend match a specific version, e.g. rename it to optee-client_4.4%.bbappend for a while, until 4.5 is out... > I agree that this patch is a good work around, but I'm not 100% sure > it's the final best fix. This way is more generic - below :remove would be no-op if such patch is not in the SRC_URI. > On 4/7/2025 10:08 PM, Denys Dmytriyenko wrote: > >From: Denys Dmytriyenko <denys@konsulko.com> > > > >meta-arm adds a patch that was sent and accepted upstream, remove > >it since mta-ti pulls a more recent verstion. > > > >Signed-off-by: Denys Dmytriyenko <denys@konsulko.com> > >--- > > .../recipes-security/optee/optee-client-ti-version.inc | 4 ++++ > > 1 file changed, 4 insertions(+) > > > >diff --git a/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc b/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc > >index 2c8977dd..a628e446 100644 > >--- a/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc > >+++ b/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc > >@@ -1,2 +1,6 @@ > > PV = "4.4.0+git" > > SRCREV = "d221676a58b305bddbf97db00395205b3038de8e" > >+ > >+SRC_URI:remove = " \ > >+ file://0001-tee-supplicant-add-udev-rule-and-systemd-service-fil.patch \ > >+"
diff --git a/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc b/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc index 2c8977dd..a628e446 100644 --- a/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc +++ b/meta-ti-bsp/recipes-security/optee/optee-client-ti-version.inc @@ -1,2 +1,6 @@ PV = "4.4.0+git" SRCREV = "d221676a58b305bddbf97db00395205b3038de8e" + +SRC_URI:remove = " \ + file://0001-tee-supplicant-add-udev-rule-and-systemd-service-fil.patch \ +"