| Message ID | 20221006234108.1696865-1-rs@ti.com |
|---|---|
| State | Accepted |
| Delegated to: | Ryan Eatmon |
| Headers | show |
| Series | [dunfell] ti-img-rogue-*: Bump the graphics recipes | expand |
On Thu, Oct 06, 2022 at 06:41:08PM -0500, Randolph Sapp via lists.yoctoproject.org wrote: > Bump the graphics recipes source revision and remove the old BVNC > switch. The BVNC is now baked into the platform Makefile that is > selected with PVR_BUILD_DIR / TARGET_PRODUCT. Any useful changelog between the 2 SRCREVs, besides removal of BVNC? > Signed-off-by: Randolph Sapp <rs@ti.com> > --- > .../ti-img-rogue-driver_1.15.6133109.bb | 10 ++-------- > .../libgles/ti-img-rogue-umlibs_1.15.6133109.bb | 2 +- > 2 files changed, 3 insertions(+), 9 deletions(-) > > diff --git a/recipes-bsp/powervr-drivers/ti-img-rogue-driver_1.15.6133109.bb b/recipes-bsp/powervr-drivers/ti-img-rogue-driver_1.15.6133109.bb > index 1127c5d1..4f6dc4d4 100644 > --- a/recipes-bsp/powervr-drivers/ti-img-rogue-driver_1.15.6133109.bb > +++ b/recipes-bsp/powervr-drivers/ti-img-rogue-driver_1.15.6133109.bb > @@ -26,7 +26,7 @@ SRC_URI = " \ > > S = "${WORKDIR}/git" > > -SRCREV = "ea2ac3fda4ac442b639bc3a0b73e9c4feac2081a" > +SRCREV = "a353b9524399d927af8e824f92902065947b2fbe" > > TARGET_PRODUCT_j7-evm = "j721e_linux" > TARGET_PRODUCT_j7-hs-evm = "j721e_linux" > @@ -34,16 +34,10 @@ TARGET_PRODUCT_j721s2-evm = "j721s2_linux" > TARGET_PRODUCT_j721s2-hs-evm = "j721s2_linux" > TARGET_PRODUCT_j784s4-evm = "j784s4_linux" > TARGET_PRODUCT_am62xx = "am62_linux" > -TARGET_BVNC_j7-evm = "22.104.208.318" > -TARGET_BVNC_j7-hs-evm = "22.104.208.318" > -TARGET_BVNC_j721s2-evm = "36.53.104.796" > -TARGET_BVNC_j721s2-hs-evm = "36.53.104.796" > -TARGET_BVNC_j784s4-evm = "36.53.104.796" > -TARGET_BVNC_am62xx = "33.15.11.3" > PVR_BUILD = "release" > PVR_WS = "wayland" > > -EXTRA_OEMAKE += 'KERNELDIR="${STAGING_KERNEL_DIR}" RGX_BVNC=${TARGET_BVNC} BUILD=${PVR_BUILD} PVR_BUILD_DIR=${TARGET_PRODUCT} WINDOW_SYSTEM=${PVR_WS}' > +EXTRA_OEMAKE += 'KERNELDIR="${STAGING_KERNEL_DIR}" BUILD=${PVR_BUILD} PVR_BUILD_DIR=${TARGET_PRODUCT} WINDOW_SYSTEM=${PVR_WS}' > > do_install() { > make -C ${STAGING_KERNEL_DIR} M=${B}/binary_${TARGET_PRODUCT}_${PVR_WS}_${PVR_BUILD}/target_aarch64/kbuild INSTALL_MOD_PATH=${D}${root_prefix} PREFIX=${STAGING_DIR_HOST} modules_install > diff --git a/recipes-graphics/libgles/ti-img-rogue-umlibs_1.15.6133109.bb b/recipes-graphics/libgles/ti-img-rogue-umlibs_1.15.6133109.bb > index b6bb4324..de979be0 100644 > --- a/recipes-graphics/libgles/ti-img-rogue-umlibs_1.15.6133109.bb > +++ b/recipes-graphics/libgles/ti-img-rogue-umlibs_1.15.6133109.bb > @@ -15,7 +15,7 @@ PR = "r2" > BRANCH = "linuxws/dunfell/k5.10/${PV}_unified_fw_pagesize" > > SRC_URI = "git://git.ti.com/git/graphics/ti-img-rogue-umlibs.git;protocol=https;branch=${BRANCH}" > -SRCREV = "4207c0ea429ec751b1ff5f831a84ceeb76ad578d" > +SRCREV = "5977e82b96028f783d39c7219f016c1faf8dc5f5" > > TARGET_PRODUCT_j7-evm = "j721e_linux" > TARGET_PRODUCT_j7-hs-evm = "j721e_linux" > -- > 2.25.1
On Mon, Oct 10 2022 at 09:44:03 PM -0400, Denys Dmytriyenko
<denis@denix.org> wrote:
> Any useful changelog between the 2 SRCREVs, besides removal of BVNC?
Yeah, actually. Two main things. The default precision for GL
fragmentation shaders was bumped up, which helps with some visual bugs
in applications like glmark2. In addition to this, a firmware patch was
added that should help some of the hardware recoveries clients were
experiencing on am62.
This is roughly the latest commit message for the selected UM version.
KM was really only a Makefile cleanup because I was tired of trying to
remember what platform had what BVNC and there were a few legacy
variables being set that had no effect.
On Tue, Oct 11, 2022 at 04:20:31PM -0500, Randolph Sapp via lists.yoctoproject.org wrote: > On Mon, Oct 10 2022 at 09:44:03 PM -0400, Denys Dmytriyenko > <denis@denix.org> wrote: > >Any useful changelog between the 2 SRCREVs, besides removal of BVNC? > > Yeah, actually. Two main things. The default precision for GL > fragmentation shaders was bumped up, which helps with some visual > bugs in applications like glmark2. In addition to this, a firmware > patch was added that should help some of the hardware recoveries > clients were experiencing on am62. > > This is roughly the latest commit message for the selected UM version. > > KM was really only a Makefile cleanup because I was tired of trying > to remember what platform had what BVNC and there were a few legacy > variables being set that had no effect. Should these be mentioned in the git commit message?
On Tue, Oct 11 2022 at 07:16:49 PM -0400, Denys Dmytriyenko
<denis@denix.org> wrote:
> Should these be mentioned in the git commit message?
Depends on what we're actually trying to track in these commit
messages. As far as I'm concerned these commit messages should only
track changes to the recipes and reasoning, but if you believe they
should also track upstream change notes then I will modify it
accordingly.
On Wed, Oct 12, 2022 at 11:32:55AM -0500, Sapp, Randolph wrote: > On Tue, Oct 11 2022 at 07:16:49 PM -0400, Denys Dmytriyenko > <denis@denix.org> wrote: > >Should these be mentioned in the git commit message? > > Depends on what we're actually trying to track in these commit > messages. As far as I'm concerned these commit messages should only > track changes to the recipes and reasoning, but if you believe they > should also track upstream change notes then I will modify it > accordingly. And that's part of the reasoning. If those upstream change notes are public, you can simply link to them in git history. Otherwise it is recommended to provide a brief high level summary of the changes in the corresponding component. https://www.openembedded.org/wiki/Commit_Patch_Message_Guidelines
Ah, alright, so should I just add the following link to the commit message then? https://git.ti.com/cgit/graphics/ti-img-rogue-umlibs/commit/?h=linuxws/dunfell/k5.10/1.15.6133109_unified_fw_pagesize&id=5977e82b96028f783d39c7219f016c1faf8dc5f5 Sorry, I didn't see a good example of this in the history. Almost all the commits I can find just indicate that they changed the SRCREV or bumped a version for an SDK release.
diff --git a/recipes-bsp/powervr-drivers/ti-img-rogue-driver_1.15.6133109.bb b/recipes-bsp/powervr-drivers/ti-img-rogue-driver_1.15.6133109.bb index 1127c5d1..4f6dc4d4 100644 --- a/recipes-bsp/powervr-drivers/ti-img-rogue-driver_1.15.6133109.bb +++ b/recipes-bsp/powervr-drivers/ti-img-rogue-driver_1.15.6133109.bb @@ -26,7 +26,7 @@ SRC_URI = " \ S = "${WORKDIR}/git" -SRCREV = "ea2ac3fda4ac442b639bc3a0b73e9c4feac2081a" +SRCREV = "a353b9524399d927af8e824f92902065947b2fbe" TARGET_PRODUCT_j7-evm = "j721e_linux" TARGET_PRODUCT_j7-hs-evm = "j721e_linux" @@ -34,16 +34,10 @@ TARGET_PRODUCT_j721s2-evm = "j721s2_linux" TARGET_PRODUCT_j721s2-hs-evm = "j721s2_linux" TARGET_PRODUCT_j784s4-evm = "j784s4_linux" TARGET_PRODUCT_am62xx = "am62_linux" -TARGET_BVNC_j7-evm = "22.104.208.318" -TARGET_BVNC_j7-hs-evm = "22.104.208.318" -TARGET_BVNC_j721s2-evm = "36.53.104.796" -TARGET_BVNC_j721s2-hs-evm = "36.53.104.796" -TARGET_BVNC_j784s4-evm = "36.53.104.796" -TARGET_BVNC_am62xx = "33.15.11.3" PVR_BUILD = "release" PVR_WS = "wayland" -EXTRA_OEMAKE += 'KERNELDIR="${STAGING_KERNEL_DIR}" RGX_BVNC=${TARGET_BVNC} BUILD=${PVR_BUILD} PVR_BUILD_DIR=${TARGET_PRODUCT} WINDOW_SYSTEM=${PVR_WS}' +EXTRA_OEMAKE += 'KERNELDIR="${STAGING_KERNEL_DIR}" BUILD=${PVR_BUILD} PVR_BUILD_DIR=${TARGET_PRODUCT} WINDOW_SYSTEM=${PVR_WS}' do_install() { make -C ${STAGING_KERNEL_DIR} M=${B}/binary_${TARGET_PRODUCT}_${PVR_WS}_${PVR_BUILD}/target_aarch64/kbuild INSTALL_MOD_PATH=${D}${root_prefix} PREFIX=${STAGING_DIR_HOST} modules_install diff --git a/recipes-graphics/libgles/ti-img-rogue-umlibs_1.15.6133109.bb b/recipes-graphics/libgles/ti-img-rogue-umlibs_1.15.6133109.bb index b6bb4324..de979be0 100644 --- a/recipes-graphics/libgles/ti-img-rogue-umlibs_1.15.6133109.bb +++ b/recipes-graphics/libgles/ti-img-rogue-umlibs_1.15.6133109.bb @@ -15,7 +15,7 @@ PR = "r2" BRANCH = "linuxws/dunfell/k5.10/${PV}_unified_fw_pagesize" SRC_URI = "git://git.ti.com/git/graphics/ti-img-rogue-umlibs.git;protocol=https;branch=${BRANCH}" -SRCREV = "4207c0ea429ec751b1ff5f831a84ceeb76ad578d" +SRCREV = "5977e82b96028f783d39c7219f016c1faf8dc5f5" TARGET_PRODUCT_j7-evm = "j721e_linux" TARGET_PRODUCT_j7-hs-evm = "j721e_linux"
Bump the graphics recipes source revision and remove the old BVNC switch. The BVNC is now baked into the platform Makefile that is selected with PVR_BUILD_DIR / TARGET_PRODUCT. Signed-off-by: Randolph Sapp <rs@ti.com> --- .../ti-img-rogue-driver_1.15.6133109.bb | 10 ++-------- .../libgles/ti-img-rogue-umlibs_1.15.6133109.bb | 2 +- 2 files changed, 3 insertions(+), 9 deletions(-)