Fix the regression from the mesa refactoring and from switching to armv7
in hammerhead.
[ci:skip-build]: already built successfully in CI for x86_64; I've
tested manually that it builds for armv7.
- main: postmarketos-mkinitfs: Add /run
Create the /run directory in the initramfs so that cryptsetup doesn't
crash on not being able to create /run/cryptsetup for lockfiles
- device: pine-dontbeevil: Remove workaround for touchscreen
- main: linux-postmarketos-allwinner: Implemented supplies for touchscreen
This uses a slightly newer commit from the kernel repo that implements
requesting a regulator in the touchscreen driver
- device: pine-dontbeevil: Remove ethernet from initramfs
This actually makes debugging more complicated since networkmanager wont
touch the ethernet adapter afterwards and the initramfs can be debugged
over uart
- temp/u-boot: sync APKBUILD with upstream and update to 2019.04
- main: linux-postmarketos-allwinner: Use 20190521 git version
Fix missing symlink for the first driver in mv_dri() (caused by "shift"
at the wrong position). Build freedreno for arm only, not for all
arches.
Lima is for rendering and sun4i-drm (part of kmsro) for the display, so
split it into its own subpackage and use it in device-pine-*.
[ci:skip-build]: ollieparanoid made sure, that this builds. This won't
finish in time in CI.
Fix up for !359: the pkgrel was not bumped, and because the real mesa
version was stored in _pkgver instead of pkgver (which is set to 9999),
this means that the APKBUILD has exactly the same version as before. So
pmbootstrap will not try to build it, resulting in a missing
mesa-dri-lima package and general confusion.
See also:
https://gitlab.com/postmarketOS/pmaports/merge_requests/359#note_170054969
temp/mesa: enable mesa-dri-lima subpackage
temp/mesa: add a fix for Wayland on Lima
temp/mesa-purism-gc7000: update to latest Git and try to resolve conflicts with mesa
temp/mesa: switch version to 19.1.0-rc1
device/device-pine-a64lts: add mesa and mesa-dri-lima to depends
device/device-pine-dontbeevil: add mesa to depends
device/device-purism-librem5dev: remove now defunct mesa-purism-gc7000-dri-imx from depends
temp/mesa*: re-add X11 support
temp/mesa-purism-gc7000: fix build error which for some reason didn't cause problems earlier
temp/mesa*: fix X11 DRI symlinks
They are available in Alpine again for armhf and armv7.
Networkmanager is still missing, so we keep it it temp for now. Upstream
patch submitted: https://github.com/alpinelinux/aports/pull/7626
Related: #244
Due to the following dependency chain, it is currently not available
in Alpine for armhf and armv7. This breaks postmarketos-base.
postmarketos-base -> networkmanager -> polkit -> mozjs60
This can be reverted, when networkmanager is back for arm{hf,v7} in
Alpine. The mozjs60 arm fix is from here:
https://github.com/alpinelinux/aports/pull/7561
Related: #244
Rebuild to fix missing dependencies:
so:libicui18n.so.63, so:libicuuc.so.63
Currently we are facing another upstream compatibility issue with
Alpine, which is why CI will fail: polkit and therefore
networkmanager don't exist for armhf, armv7. See #244 for details.
Work around for issue #1715
Link qemu static binaries to jemalloc. This malloc implementation does not
appear to be susceptible to the deadlock after fork issue.
This is package is meant to work around the qemu static deadlock while
cross-compiling.
- pmbootstrap binfmt_misc chroots on arm hosts are untested
- upstream jemalloc package has armv7 disabled. It is uncertain what is
wrong with jemalloc for armv7.
For these reasons, disable the arm packages.
This updates the PKGBUILD to build tags instead of the release branch
(which will see commits until the specific release is EOL),
This commit also adds building the mesa-xatracker package.
[ci:skip-build]: already built successfully in CI
This adds a Mesa package overriding Mesa in Alpine, and:
1) builds a much newer version of Mesa than what is found in Alpine (by
*not* building amd/radeonsi support, so no dependency on llvm).
2) builds Mesa with tegra and etnaviv drivers (note, as of now, the
librem5 devkit gpu is not supported in the etnaviv driver here)
This breaks "pmbootstrap qemu" for some people, use --host-qemu as
workaround for now (a fix will be available soon):
https://wiki.postmarketos.org/wiki/Troubleshooting:QEMU#Using_QEMU_from_the_host_system
Since this also fixes the blackscreen bug on lg-hammerhead, the
trade-off with the qemu bug is worth it.
[ci:skip-build]: already built successfully in CI
This adds a package overriding the libdrm in Alpine with a newer
version, and also building libdrm with support for experimental devices
(etnaviv and tegra).
Fork the package from Alpine and undefine IFA_F_NOPREFIXROUTE during
compilation, so it works with the old downstream kernels that do not
have this feature yet.
Plasma-framework 5.55.0 isn't building for aarch64 without this patch.
Aleix Pol ran into the same problem with flatpak builds, and created the
patch: https://codereview.qt-project.org/#/c/253276/
* Mir starts up and is able to display system settings
* x86_64 only for now, because at least ubuntu-app-test did not build
on aarch64
Based on PureTryOut's work. Getting it to this stage was a huge effort
(as it shows in the package count: 111(!)). See the merge request for
details.
[skip ci]: this won't finish in CI; ollieparanoid made sure that
everything builds for x86_64.
Alpine Linux has updated their main/icu to 63.1, which means we have to
rebuild some of our packages too.
[skip ci]: simple pkgrel bumps, building won't finish in CI
Turns out we do need to fork this aport, as it builds qgpgme, which
Alpine's aport does not build as of now. Updated to 1.12.0 and changed
the pkgver to 9999, so it does not conflict with Alpine's aport.
We had probably added this, because no binary package was present in
Alpine at the time. Now it is present, and the package has been updated
in Alpine. No need to keep this outdated version around.
Workaround for a build error on aarch64, where the binary package
repository is currently stuck. Compiling it for aarch64 like this:
$ pmbootstrap build --strict --arch=aarch64 qemu
Resulted in:
>>> qemu: Analyzing dependencies...
ERROR: unsatisfiable constraints:
openssl-dev-1.1.1a-r0:
conflicts: libressl-dev-2.7.4-r2[pc:libcrypto=1.1.1a]
libressl-dev-2.7.4-r2[pc:libssl=1.1.1a]
libressl-dev-2.7.4-r2[pc:openssl=1.1.1a]
satisfies: curl-dev-7.62.0-r2[openssl-dev]
libssh2-dev-1.8.0-r4[pc:libcrypto]
libssh2-dev-1.8.0-r4[pc:libssl]
spice-dev-0.14.1-r3[pc:openssl]
libressl-dev-2.7.4-r2:
conflicts: openssl-dev-1.1.1a-r0[pc:libcrypto=2.7.4]
openssl-dev-1.1.1a-r0[pc:libssl=2.7.4]
openssl-dev-1.1.1a-r0[pc:openssl=2.7.4]
satisfies: world[libressl-dev]
libssh2-dev-1.8.0-r4[pc:libcrypto]
libssh2-dev-1.8.0-r4[pc:libssl]
spice-dev-0.14.1-r3[pc:openssl]
>>> ERROR: qemu: builddeps failed
Note that the only package not mentioned in both "satisfies" outputs is
curl-dev. The real questions are: why is libressl-dev getting pulled in
at all? (Alpine switched back from libressl to openssl, so this should
not happen). And why does this only happen for aarch64, but not for
x86_64 and armhf? But at least this patch unblocks the package builder.
