Display manager proposal for x11 and wayland
v6 cover Display manager proposal for x11 and wayland
From: Randolph Sapp <[email protected]>
We've recently run into some issues with weston-init attempting to start Weston prior to all drm devices being registered. There's not really a good, scriptable mechanism to listen in to device registration events that works with the existing weston-init package. Well, at least one that doesn't involve polling files or introducing more dependency on the init system being used.
I also see there is also a lot of scripting around starting X11, xserver-nodm-init, that (from my limited review) should experience the same issue.
I'd like to introduce the following display manager for oe-core, emptty [1]. This display manager is, as described upstream, a "Dead simple CLI Display Manager on TTY". It supports both x11 and wayland sessions, with togglable build parameters to completely remove x11 and pam dependencies. It's licensed MIT, which shouldn't be an issue for any users. (It is written in Go, if you have opinions about that.)
With this, both weston-init and the xserver-nodm-init packages can be re-tuned to leverage this display manager and simply add a user and emptty config for an autologin session. This can resolve the current behavior across init systems without additional scripting, and move some development out of this layer.
This lists myself as a maintainer of emptty as well as xserver-nodm-init and xuser-account since these are currently unassigned and I've reworked them significantly here.
Sorry for the delay on this series. I found a few bugs in emptty that I wanted to address before submitting this officially.
[1] https://github.com/tvrzna/emptty
v2:
- Address spelling issues in commit messages
- Attempt to resolve some test related issues with weston
- Add additional logs to X11 related tests
v3:
- Reset AUTOLOGIN_MAX_RETRY to the default value of 2. When running under QEMU the first auth attempt almost always fails.
v4:
- Add a tmpfile entry for the x11 domain socket directory.
- Remove some scripts associated with weston-init that were being shipped with weston
v5:
- Move tmpfile data to individual files
- Add explicit entries for these in the FILES variable
v6:
- Do not attempt to ship a tmpfiles.d entry in libx11
Randolph Sapp (5):
libx11: create tmpfile dir for x11 domain socket
emptty: add version 0.14.0
weston-init: convert to virtual-emptty-conf
weston: remove deprecated weston-start scripts
xserver-nodm-init: convert to virtual-emptty-conf
.../conf/distro/include/default-providers.inc | 1 + meta/conf/distro/include/maintainers.inc | 6 +- meta/lib/oeqa/runtime/cases/weston.py | 18 +- meta/lib/oeqa/runtime/cases/xorg.py | 8 + meta/recipes-graphics/emptty/emptty-conf.bb | 14 + meta/recipes-graphics/emptty/emptty.inc | 26 ++ .../recipes-graphics/emptty/emptty/emptty.tab | 1 + meta/recipes-graphics/emptty/emptty/pamconf | 10 + meta/recipes-graphics/emptty/emptty_0.14.0.bb | 53 +++ meta/recipes-graphics/wayland/weston-init.bb | 61 +-- .../wayland/weston-init/emptty.conf | 77 ++++ .../recipes-graphics/wayland/weston-init/init | 54 --- .../wayland/weston-init/weston-autologin | 11 - .../wayland/weston-init/weston-socket.sh | 20 - .../wayland/weston-init/weston-start | 76 ---- .../wayland/weston-init/weston.env | 0 .../wayland/weston-init/weston.service | 71 ---- .../wayland/weston-init/weston.socket | 14 - .../weston/systemd-notify.weston-start | 9 - .../wayland/weston/xwayland.weston-start | 6 - .../recipes-graphics/wayland/weston_14.0.1.bb | 10 - .../x11-common/xserver-nodm-init/X11/Xsession | 38 -- .../X11/Xsession.d/13xdgbasedirs.sh | 19 - .../X11/Xsession.d/89xdgautostart.sh | 7 - .../X11/Xsession.d/90XWindowManager.sh | 7 - .../x11-common/xserver-nodm-init/Xserver | 25 -- .../xserver-nodm-init/capability.conf | 2 - .../xserver-nodm-init/default.desktop | 5 + .../xserver-nodm-init/emptty.conf.in | 77 ++++ .../xserver-nodm-init/gplv2-license.patch | 355 ------------------ .../x11-common/xserver-nodm-init/xserver-nodm | 75 ---- .../xserver-nodm-init/xserver-nodm.conf.in | 7 - .../xserver-nodm-init/xserver-nodm.service.in | 11 - .../x11-common/xserver-nodm-init_3.0.bb | 57 +-- meta/recipes-graphics/xorg-lib/libx11/99_x11 | 1 + .../xorg-lib/libx11_1.8.12.bb | 15 +- .../user-creation/xuser-account_0.1.bb | 3 +- 37 files changed, 332 insertions(+), 918 deletions(-) create mode 100644 meta/recipes-graphics/emptty/emptty-conf.bb create mode 100644 meta/recipes-graphics/emptty/emptty.inc create mode 100644 meta/recipes-graphics/emptty/emptty/emptty.tab create mode 100644 meta/recipes-graphics/emptty/emptty/pamconf create mode 100644 meta/recipes-graphics/emptty/emptty_0.14.0.bb create mode 100644 meta/recipes-graphics/wayland/weston-init/emptty.conf delete mode 100644 meta/recipes-graphics/wayland/weston-init/init delete mode 100644 meta/recipes-graphics/wayland/weston-init/weston-autologin delete mode 100755 meta/recipes-graphics/wayland/weston-init/weston-socket.sh delete mode 100755 meta/recipes-graphics/wayland/weston-init/weston-start delete mode 100644 meta/recipes-graphics/wayland/weston-init/weston.env delete mode 100644 meta/recipes-graphics/wayland/weston-init/weston.service delete mode 100644 meta/recipes-graphics/wayland/weston-init/weston.socket delete mode 100644 meta/recipes-graphics/wayland/weston/systemd-notify.weston-start delete mode 100644 meta/recipes-graphics/wayland/weston/xwayland.weston-start delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession.d/13xdgbasedirs.sh delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession.d/89xdgautostart.sh delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession.d/90XWindowManager.sh delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/Xserver delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/capability.conf create mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/default.desktop create mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/emptty.conf.in delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/gplv2-license.patch delete mode 100755 meta/recipes-graphics/x11-common/xserver-nodm-init/xserver-nodm delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/xserver-nodm.conf.in delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/xserver-nodm.service.in create mode 100644 meta/recipes-graphics/xorg-lib/libx11/99_x11
Signature · 1 line
2.49.0
PATCH v6 1/5 libx11: create tmpfile dir for x11 domain socket
From: Randolph Sapp <[email protected]>
Register a volatile directory entry for the x11 domain socket when not using systemd. This will make sure the directory is always created with the correct permissions. Systemd already provides their own tmpfile.d entry for the same behavior.
Currently some x11 related applications will create this directory if it doesn't already exist, but this is not true for everything. In addition, if the application in question isn't started as root, it's possible this directory can be owned by a non-root user. This isn't an issue by itself, but it can potentially lead to problems in a multi-user environment.
--- meta/recipes-graphics/xorg-lib/libx11/99_x11 | 1 + meta/recipes-graphics/xorg-lib/libx11_1.8.12.bb | 15 +++++++++++++-- 2 files changed, 14 insertions(+), 2 deletions(-) create mode 100644 meta/recipes-graphics/xorg-lib/libx11/99_x11
meta/recipes-graphics/xorg-lib/libx11/99_x11 new file +1 -0
@@ -0,0 +1 @@ +d root root 1777 /tmp/.X11-unix none
meta/recipes-graphics/xorg-lib/libx11_1.8.12.bb +13 -2
@@ -22,7 +22,8 @@ PE = "1" XORG_PN = "libX11" -SRC_URI += "file://disable_tests.patch" +SRC_URI += "file://disable_tests.patch \ + file://99_x11" SRC_URI[sha256sum] = "fa026f9bb0124f4d6c808f9aef4057aad65e7b35d8ff43951cef0abe06bb9a9a"
@@ -36,7 +37,17 @@ PACKAGECONFIG[xcms] = "--enable-xcms,--disable-xcms" PACKAGES =+ "${PN}-xcb" -FILES:${PN} += "${datadir}/X11/XKeysymDB ${datadir}/X11/XErrorDB ${datadir}/X11/Xcms.txt" +do_install:append() { + # temporary directory required for x11 domain sockets + # systemd provides their own definition using tmpfiles.d + if ${@oe.utils.conditional('VIRTUAL-RUNTIME_init_manager', 'systemd', 'false', 'true', d)}; then + install -d ${D}${sysconfdir}/default/volatiles + install -m 0644 ${UNPACKDIR}/99_x11 ${D}${sysconfdir}/default/volatiles/99_x11 + fi +} + +FILES:${PN} += "${datadir}/X11/XKeysymDB ${datadir}/X11/XErrorDB ${datadir}/X11/Xcms.txt \ + ${libdir}/tmpfiles.d/x11.conf ${sysconfdir}/default/volatiles/99_x11" FILES:${PN}-xcb += "${libdir}/libX11-xcb.so.*" FILES:${PN}-locale += "${datadir}/X11/locale ${libdir}/X11/locale"
Signature · 1 line
2.49.0
PATCH v6 2/5 emptty: add version 0.14.0
In reply to [email protected]From: Randolph Sapp <[email protected]>
Add emptty, a "Dead simple CLI Display Manager on TTY". This is a relatively lightweight display manager that supports x11 as well as wayland through both a CLI and an automatic login mechanism.
This can effectively replace the custom init scripts for both x11 (xserver-nodm-init) and wayland (weston-init) with a single tool with more verbose logging capabilities.
This is split into two recipes. The emptty package provides the binary and various agnostic configs while the emptty-conf package provides the default configuration for the application itself. This RPROVIDES virtual-emptty-conf, allowing other recipes to provide their own configurations to override the default behavior.
--- .../conf/distro/include/default-providers.inc | 1 + meta/conf/distro/include/maintainers.inc | 2 + meta/recipes-graphics/emptty/emptty-conf.bb | 14 +++++ meta/recipes-graphics/emptty/emptty.inc | 26 +++++++++ .../recipes-graphics/emptty/emptty/emptty.tab | 1 + meta/recipes-graphics/emptty/emptty/pamconf | 10 ++++ meta/recipes-graphics/emptty/emptty_0.14.0.bb | 53 +++++++++++++++++++ 7 files changed, 107 insertions(+) create mode 100644 meta/recipes-graphics/emptty/emptty-conf.bb create mode 100644 meta/recipes-graphics/emptty/emptty.inc create mode 100644 meta/recipes-graphics/emptty/emptty/emptty.tab create mode 100644 meta/recipes-graphics/emptty/emptty/pamconf create mode 100644 meta/recipes-graphics/emptty/emptty_0.14.0.bb
meta/conf/distro/include/default-providers.inc +1 -0
@@ -68,3 +68,4 @@ PREFERRED_PROVIDER_nativesdk-mesa ?= "nativesdk-mesa" PREFERRED_PROVIDER_virtual/nativesdk-libsdl2 ?= "nativesdk-libsdl2" PREFERRED_RPROVIDER_virtual-x-terminal-emulator ?= "rxvt-unicode" +PREFERRED_RPROVIDER_virtual-emptty-conf ?= "emptty-conf"
meta/conf/distro/include/maintainers.inc +2 -0
@@ -172,6 +172,8 @@ RECIPE_MAINTAINER:pn-efivar = "Ross Burton <[email protected]>" RECIPE_MAINTAINER:pn-efibootmgr = "Ross Burton <[email protected]>" RECIPE_MAINTAINER:pn-elfutils = "Zang Ruochen <[email protected]>" RECIPE_MAINTAINER:pn-ell = "Zang Ruochen <[email protected]>" +RECIPE_MAINTAINER:pn-emptty = "Randolph Sapp <[email protected]>" +RECIPE_MAINTAINER:pn-emptty-conf = "Randolph Sapp <[email protected]>" RECIPE_MAINTAINER:pn-enchant2 = "Anuj Mittal <[email protected]>" RECIPE_MAINTAINER:pn-encodings = "Unassigned <[email protected]>" RECIPE_MAINTAINER:pn-epiphany = "Unassigned <[email protected]>"
meta/recipes-graphics/emptty/emptty-conf.bb new file +14 -0
@@ -0,0 +1,14 @@ +require emptty.inc + +SUMMARY += " (Default config)" + +do_configure[noexec] = "1" +do_compile[noexec] = "1" + +do_install () { + oe_runmake -C ${S}/src/${GO_IMPORT} DESTDIR=${D} install-config +} + +FILES:${PN}-conf = "${sysconfdir}/emptty/conf" +CONFFILES:${PN}-conf = "${sysconfdir}/emptty/conf" +RPROVIDES:${PN}-conf += "virtual-emptty-conf"
meta/recipes-graphics/emptty/emptty.inc new file +26 -0
@@ -0,0 +1,26 @@ +SUMMARY = "Dead simple CLI Display Manager on TTY" +DESCRIPTION = "Emptty is a simple display manager with a command line greeter \ +It supports both X11 and Wayland sessions, exporting required variables and \ +creating user paths as necessary." + +HOMEPAGE = "https://github.com/tvrzna/emptty" +BUGTRACKER = "https://github.com/tvrzna/emptty/issues" +LICENSE = "MIT" +LIC_FILES_CHKSUM = "file://src/${GO_IMPORT}/LICENSE;md5=d1e4d12c7d1d17367ba5668706a405ba" + +S = "${UNPACKDIR}" + +FILESEXTRAPATHS:append = ":${THISDIR}/emptty" + +GO_IMPORT = "github.com/tvrzna/emptty" +GO_IMPORT_pam = "github.com/msteinert/pam" +SRC_URI = "\ + git://${GO_IMPORT}.git;protocol=https;branch=master;destsuffix=src/${GO_IMPORT} \ + git://${GO_IMPORT_pam}.git;protocol=https;branch=master;name=pam;destsuffix=src/${GO_IMPORT_pam} \ + file://pamconf \ + file://emptty.tab \ + " +SRCREV = "d162bef75fadc8ae1d8e33dbb3de5e2795a196e0" +SRCREV_pam = "50ded1b0e7864b9bf75005eb945a8ec826bcf69d" + +SRCREV_FORMAT .= "_pam"
meta/recipes-graphics/emptty/emptty/emptty.tab new file +1 -0
@@ -0,0 +1 @@ +7:5:respawn:/usr/bin/emptty -t 7 -d
meta/recipes-graphics/emptty/emptty/pamconf new file +10 -0
@@ -0,0 +1,10 @@ +#%PAM-1.0 +auth sufficient pam_succeed_if.so user ingroup nopasswdlogin +auth include common-auth +-auth optional pam_gnome_keyring.so +-auth optional pam_kwallet5.so +account include common-account +session include common-session +-session optional pam_gnome_keyring.so auto_start +-session optional pam_kwallet5.so auto_start force_run +password include common-password
meta/recipes-graphics/emptty/emptty_0.14.0.bb new file +53 -0
53 lines not on this page · raw message
Signature · 1 line
2.49.0
PATCH v6 3/5 weston-init: convert to virtual-emptty-conf
In reply to [email protected]From: Randolph Sapp <[email protected]>
Convert this script package into a simple config for emptty using the same weston user but leveraging the nopasswdlogin session provided by emptty. Runtime provide virtual-emptty-conf as we need to set parameters for the default session.
--- meta/lib/oeqa/runtime/cases/weston.py | 18 +++-- meta/recipes-graphics/wayland/weston-init.bb | 61 +++------------ .../wayland/weston-init/emptty.conf | 77 +++++++++++++++++++ .../recipes-graphics/wayland/weston-init/init | 54 ------------- .../wayland/weston-init/weston-autologin | 11 --- .../wayland/weston-init/weston-socket.sh | 20 ----- .../wayland/weston-init/weston-start | 76 ------------------ .../wayland/weston-init/weston.env | 0 .../wayland/weston-init/weston.service | 71 ----------------- .../wayland/weston-init/weston.socket | 14 ---- 10 files changed, 101 insertions(+), 301 deletions(-) create mode 100644 meta/recipes-graphics/wayland/weston-init/emptty.conf delete mode 100644 meta/recipes-graphics/wayland/weston-init/init delete mode 100644 meta/recipes-graphics/wayland/weston-init/weston-autologin delete mode 100755 meta/recipes-graphics/wayland/weston-init/weston-socket.sh delete mode 100755 meta/recipes-graphics/wayland/weston-init/weston-start delete mode 100644 meta/recipes-graphics/wayland/weston-init/weston.env delete mode 100644 meta/recipes-graphics/wayland/weston-init/weston.service delete mode 100644 meta/recipes-graphics/wayland/weston-init/weston.socket
meta/lib/oeqa/runtime/cases/weston.py +12 -6
@@ -37,7 +37,7 @@ class WestonTest(OERuntimeTestCase): def run_weston_init(self): if 'systemd' in self.tc.td['VIRTUAL-RUNTIME_init_manager']: - self.target.run('systemd-run --collect --unit=weston-ptest.service --uid=0 -p PAMName=login -p TTYPath=/dev/tty6 -E XDG_RUNTIME_DIR=/tmp -E WAYLAND_DISPLAY=wayland-0 /usr/bin/weston --socket=wayland-1 --log=%s' % self.weston_log_file) + self.target.run(self.get_weston_command('systemd-run --collect --unit=weston-ptest.service --uid=0 -p PAMName=login -p TTYPath=/dev/tty6 -E XDG_RUNTIME_DIR -E WAYLAND_DISPLAY /usr/bin/weston --socket=wayland-2 --log=%s' % self.weston_log_file)) else: self.target.run(self.get_weston_command('openvt -- weston --socket=wayland-2 --log=%s' % self.weston_log_file))
@@ -55,12 +55,18 @@ class WestonTest(OERuntimeTestCase): @OEHasPackage(['wayland-utils']) def test_wayland_info(self): - if 'systemd' in self.tc.td['VIRTUAL-RUNTIME_init_manager']: - command = 'XDG_RUNTIME_DIR=/run wayland-info' - else: - command = self.get_weston_command('wayland-info') + command = self.get_weston_command('wayland-info') status, output = self.target.run(command) - self.assertEqual(status, 0, msg='wayland-info error: %s' % output) + msg = 'wayland-info error: %s' % output + + # dump last 20 lines of emptty log in case of failure + log_cmd = 'tail -n 20 /var/log/emptty/7.log' + msg += '\n\n===== start: snippet =====\n\n' + msg += 'file: /var/log/emptty/7.log\n\n' + msg += '\n\n%s\n\n' % self.target.run(log_cmd)[1] + msg += '\n\n===== end: snippet =====\n\n' + + self.assertEqual(status, 0, msg=msg) @OEHasPackage(['weston']) def test_weston_can_initialize_new_wayland_compositor(self):
meta/recipes-graphics/wayland/weston-init.bb +12 -49
93 lines not on this page · raw message
meta/recipes-graphics/wayland/weston-init/emptty.conf new file +77 -0
77 lines not on this page · raw message
meta/recipes-graphics/wayland/weston-init/init deleted +0 -54
54 lines not on this page · raw message
meta/recipes-graphics/wayland/weston-init/weston-autologin deleted +0 -11
@@ -1,11 +0,0 @@ -auth required pam_nologin.so -auth required pam_unix.so try_first_pass nullok - -account required pam_nologin.so -account required pam_unix.so - -session required pam_env.so -session required pam_unix.so --session optional pam_systemd.so type=wayland class=user desktop=weston --session optional pam_loginuid.so -
meta/recipes-graphics/wayland/weston-init/weston-socket.sh deleted +0 -20
@@ -1,20 +0,0 @@ -#!/bin/sh - -# set weston variables for use with global weston socket -global_socket="/run/wayland-0" -if [ -e "$global_socket" ]; then - weston_group=$(stat -c "%G" "$global_socket") - if [ "$(id -u)" = "0" ]; then - export WAYLAND_DISPLAY="$global_socket" - else - case "$(groups "$USER")" in - *"$weston_group"*) - export WAYLAND_DISPLAY="$global_socket" - ;; - *) - ;; - esac - fi - unset weston_group -fi -unset global_socket
meta/recipes-graphics/wayland/weston-init/weston-start deleted +0 -76
76 lines not on this page · raw message
meta/recipes-graphics/wayland/weston-init/weston.env deleted +0 -0
meta/recipes-graphics/wayland/weston-init/weston.service deleted +0 -71
71 lines not on this page · raw message
meta/recipes-graphics/wayland/weston-init/weston.socket deleted +0 -14
@@ -1,14 +0,0 @@ -[Unit] -Description=Weston socket -RequiresMountsFor=/run - -[Socket] -ListenStream=/run/wayland-0 -SocketMode=0775 -SocketUser=weston -SocketGroup=wayland -RemoveOnStop=yes - -[Install] -WantedBy=sockets.target -
Signature · 1 line
2.49.0
PATCH v6 4/5 weston: remove deprecated weston-start scripts
In reply to [email protected]From: Randolph Sapp <[email protected]>
The weston-start scripts are not parsed unless using the old version of weston-init. These files are no longer required since weston-init moved to a virtual-emptty-conf.
--- .../wayland/weston/systemd-notify.weston-start | 9 --------- .../wayland/weston/xwayland.weston-start | 6 ------ meta/recipes-graphics/wayland/weston_14.0.1.bb | 10 ---------- 3 files changed, 25 deletions(-) delete mode 100644 meta/recipes-graphics/wayland/weston/systemd-notify.weston-start delete mode 100644 meta/recipes-graphics/wayland/weston/xwayland.weston-start
meta/recipes-graphics/wayland/weston/systemd-notify.weston-start deleted +0 -9
@@ -1,9 +0,0 @@ -#!/bin/sh - -# SPDX-FileCopyrightText: Huawei Inc. -# SPDX-License-Identifier: Apache-2.0 - - -if [[ -x "/usr/lib/weston/systemd-notify.so" ]]; then - add_weston_module "systemd-notify.so" -fi
meta/recipes-graphics/wayland/weston/xwayland.weston-start deleted +0 -6
@@ -1,6 +0,0 @@ -#!/bin/sh - -if type Xwayland >/dev/null 2>/dev/null; then - mkdir -m 775 -p /tmp/.X11-unix - chown root:video /tmp/.X11-unix -fi
meta/recipes-graphics/wayland/weston_14.0.1.bb +0 -10
@@ -9,8 +9,6 @@ LIC_FILES_CHKSUM = "file://COPYING;md5=d79ee9e66bb0f95d3386a7acae780b70 \ SRC_URI = "https://gitlab.freedesktop.org/wayland/weston/-/releases/${PV}/downloads/${BPN}-${PV}.tar.xz \ file://weston.png \ file://weston.desktop \ - file://xwayland.weston-start \ - file://systemd-notify.weston-start \ " SRC_URI[sha256sum] = "a8150505b126a59df781fe8c30c8e6f87da7013e179039eb844a5bbbcc7c79b3"
@@ -108,14 +106,6 @@ do_install:append() { install ${UNPACKDIR}/weston.png ${D}${datadir}/icons/hicolor/48x48/apps fi - if [ "${@bb.utils.contains('PACKAGECONFIG', 'xwayland', 'yes', 'no', d)}" = "yes" ]; then - install -Dm 644 ${UNPACKDIR}/xwayland.weston-start ${D}${datadir}/weston-start/xwayland - fi - - if [ "${@bb.utils.contains('PACKAGECONFIG', 'systemd', 'yes', 'no', d)}" = "yes" ]; then - install -Dm 644 ${UNPACKDIR}/systemd-notify.weston-start ${D}${datadir}/weston-start/systemd-notify - fi - if [ "${@bb.utils.contains('PACKAGECONFIG', 'launch', 'yes', 'no', d)}" = "yes" ]; then chmod u+s ${D}${bindir}/weston-launch fi
Signature · 1 line
2.49.0
PATCH v6 5/5 xserver-nodm-init: convert to virtual-emptty-conf
In reply to [email protected]From: Randolph Sapp <[email protected]>
Convert this script package into a simple config for emptty. This is capable of using either the root user or the underprivileged xuser account based on the same ROOTLESS_X variable. The xuser will leverage the nopasswdlogin session provided by emptty for automatic logins. Runtime provide virtual-emptty-conf as we need to set parameters for the default session.
This also gets rid of the xuser security/capability override since emptty sets up required components as root before decending into the specified user. This does not currently handle loading of autostart applications in "/etc/xdg/autostart".
License-Update: GPLv2 scripts were replaced with MIT based config files
---
meta/conf/distro/include/maintainers.inc | 4 +-
meta/lib/oeqa/runtime/cases/xorg.py | 8 +
.../x11-common/xserver-nodm-init/X11/Xsession | 38 --
.../X11/Xsession.d/13xdgbasedirs.sh | 19 -
.../X11/Xsession.d/89xdgautostart.sh | 7 -
.../X11/Xsession.d/90XWindowManager.sh | 7 -
.../x11-common/xserver-nodm-init/Xserver | 25 --
.../xserver-nodm-init/capability.conf | 2 -
.../xserver-nodm-init/default.desktop | 5 +
.../xserver-nodm-init/emptty.conf.in} | 8 +-
.../xserver-nodm-init/gplv2-license.patch | 355 ------------------
.../x11-common/xserver-nodm-init/xserver-nodm | 75 ----
.../xserver-nodm-init/xserver-nodm.conf.in | 7 -
.../xserver-nodm-init/xserver-nodm.service.in | 11 -
.../x11-common/xserver-nodm-init_3.0.bb | 57 +--
.../user-creation/xuser-account_0.1.bb | 3 +-
16 files changed, 37 insertions(+), 594 deletions(-)
delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession
delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession.d/13xdgbasedirs.sh
delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession.d/89xdgautostart.sh
delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession.d/90XWindowManager.sh
delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/Xserver
delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/capability.conf
create mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/default.desktop
copy meta/recipes-graphics/{wayland/weston-init/emptty.conf => x11-common/xserver-nodm-init/emptty.conf.in} (95%)
delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/gplv2-license.patch
delete mode 100755 meta/recipes-graphics/x11-common/xserver-nodm-init/xserver-nodm
delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/xserver-nodm.conf.in
delete mode 100644 meta/recipes-graphics/x11-common/xserver-nodm-init/xserver-nodm.service.inmeta/conf/distro/include/maintainers.inc +2 -2
@@ -921,12 +921,12 @@ RECIPE_MAINTAINER:pn-xorgproto = "Unassigned <[email protected]>" RECIPE_MAINTAINER:pn-xprop = "Unassigned <[email protected]>" RECIPE_MAINTAINER:pn-xrandr = "Unassigned <[email protected]>" RECIPE_MAINTAINER:pn-xrestop = "Unassigned <[email protected]>" -RECIPE_MAINTAINER:pn-xserver-nodm-init = "Unassigned <[email protected]>" +RECIPE_MAINTAINER:pn-xserver-nodm-init = "Randolph Sapp <[email protected]>" RECIPE_MAINTAINER:pn-xserver-xf86-config = "Unassigned <[email protected]>" RECIPE_MAINTAINER:pn-xserver-xorg = "Unassigned <[email protected]>" RECIPE_MAINTAINER:pn-xset = "Unassigned <[email protected]>" RECIPE_MAINTAINER:pn-xtrans = "Unassigned <[email protected]>" -RECIPE_MAINTAINER:pn-xuser-account = "Unassigned <[email protected]>" +RECIPE_MAINTAINER:pn-xuser-account = "Randolph Sapp <[email protected]>" RECIPE_MAINTAINER:pn-xvinfo = "Unassigned <[email protected]>" RECIPE_MAINTAINER:pn-xwayland = "Unassigned <[email protected]>" RECIPE_MAINTAINER:pn-xwininfo = "Unassigned <[email protected]>"
meta/lib/oeqa/runtime/cases/xorg.py +8 -0
@@ -20,4 +20,12 @@ class XorgTest(OERuntimeTestCase): status, output = self.target.run(cmd) msg = ('Xorg does not appear to be running %s' % self.target.run(self.tc.target_cmds['ps'])[1]) + + # dump last 20 lines of emptty log in case of failure + log_cmd = 'tail -n 20 /var/log/emptty/7.log' + msg += '\n\n===== start: snippet =====\n\n' + msg += 'file: /var/log/emptty/7.log\n\n' + msg += '\n\n%s\n\n' % self.target.run(log_cmd)[1] + msg += '\n\n===== end: snippet =====\n\n' + self.assertEqual(status, 0, msg=msg)
meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession deleted +0 -38
@@ -1,38 +0,0 @@ -#!/bin/sh - -if [ -x /usr/bin/dbus-launch ]; then - # As this is the X session script, always start a new DBus session. - eval `dbus-launch --sh-syntax --exit-with-session </dev/null` - echo "D-BUS per-session daemon address is: $DBUS_SESSION_BUS_ADDRESS" -fi - -. /etc/profile - -if [ -f $HOME/.profile ]; then - . $HOME/.profile -fi - -SYSSESSIONDIR=/etc/X11/Xsession.d - -export CLUTTER_DISABLE_MIPMAPPED_TEXT=1 - -for SESSIONFILE in $SYSSESSIONDIR/*; do - set +e - case "$SESSIONFILE" in - *.sh) - . "$SESSIONFILE" - ;; - *.shbg) - "$SESSIONFILE" & - ;; - *~) - # Ignore backup files - ;; - *) - "$SESSIONFILE" - ;; - esac - set -e -done - -exit 0
meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession.d/13xdgbasedirs.sh deleted +0 -19
@@ -1,19 +0,0 @@ -# Minimal/stub implementation of the XDG Base Directory specification. -# http://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html - -# If the runtime directory hasn't been set already (for example by systemd, -# elogind, or pam) create a directory in TMPDIR. -if [ -z "$XDG_RUNTIME_DIR" ]; then - XDG_RUNTIME_DIR=/run/user/$(id -u) - export XDG_RUNTIME_DIR -fi - -if [ -d "$XDG_RUNTIME_DIR" ]; then - # If the directory exists, check the permissions and ownership - if [ "$(stat -c %u-%a "$XDG_RUNTIME_DIR")" != "$(id -u)-700" ]; then - echo "ERROR: $XDG_RUNTIME_DIR has incorrect permissions" - exit 1 - fi -else - mkdir --mode 0700 --parents "${XDG_RUNTIME_DIR}" -fi
meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession.d/89xdgautostart.sh deleted +0 -7
@@ -1,7 +0,0 @@ -XDGAUTOSTART=/etc/xdg/autostart -if [ -d $XDGAUTOSTART ]; then - for SCRIPT in $XDGAUTOSTART/*; do - CMD=`grep ^Exec= $SCRIPT | cut -d '=' -f 2` - $CMD & - done -fi
meta/recipes-graphics/x11-common/xserver-nodm-init/X11/Xsession.d/90XWindowManager.sh deleted +0 -7
@@ -1,7 +0,0 @@ -if [ -x $HOME/.Xsession ]; then - exec $HOME/.Xsession -elif [ -x /usr/bin/x-session-manager ]; then - exec /usr/bin/x-session-manager -else - exec /usr/bin/x-window-manager -fi
meta/recipes-graphics/x11-common/xserver-nodm-init/Xserver deleted +0 -25
@@ -1,25 +0,0 @@ -#!/bin/sh - -# This script is only needed to make sure /etc/X11/xserver-common -# can affect XSERVER, ARGS & DPI: otherwise systemd could just use -# /etc/default/xserver-nodm as EnvironmentFile and sysvinit could just -# source the same file - -. /etc/profile - -# load default values for XSERVER, ARGS, DISPLAY... -. /etc/default/xserver-nodm - -# Allow xserver-common to override ARGS, XSERVER, DPI -if [ -e /etc/X11/xserver-common ] ; then - . /etc/X11/xserver-common - if [ ! -e $XSERVER ] ; then - XSERVER=$(which $XSERVER) - fi -fi - -if [ -n "$DPI" ] ; then - ARGS="$ARGS -dpi $DPI" -fi - -exec xinit /etc/X11/Xsession -- $XSERVER $DISPLAY $ARGS $*
meta/recipes-graphics/x11-common/xserver-nodm-init/capability.conf deleted +0 -2
@@ -1,2 +0,0 @@ -cap_sys_admin @USER@ -none *
meta/recipes-graphics/x11-common/xserver-nodm-init/default.desktop new file +5 -0
@@ -0,0 +1,5 @@ +[Desktop Entry] +Encoding=UTF-8 +Type=Application +Name=Default X11 Desktop +Exec=x-session-manager
meta/recipes-graphics/x11-common/xserver-nodm-init/emptty.conf.in renamed from meta/recipes-graphics/wayland/weston-init/emptty.conf +4 -4
@@ -11,13 +11,13 @@ PRINT_ISSUE=true PRINT_MOTD=true # Preselected user, if AUTOLOGIN is enabled, this user is logged in. -DEFAULT_USER=weston +DEFAULT_USER=@USER@ # Enables Autologin, if DEFAULT_USER is defined and part of nopasswdlogin group. Possible values are "true" or "false". AUTOLOGIN=true # The default session used, if Autologin is enabled. If session is not found in list of session, it proceeds to manual selection. -AUTOLOGIN_SESSION=Weston +AUTOLOGIN_SESSION=Default X11 Desktop # If Autologin is enabled and session does not start correctly, the number of retries in short period is kept to eventually stop the infinite loop of restarts. -1 is for infinite retries, 0 is for no retry. # AUTOLOGIN_MAX_RETRY=2
@@ -41,7 +41,7 @@ VERTICAL_SELECTION=false #LOGGING_FILE=/var/log/emptty/[TTY_NUMBER].log # Arguments passed to Xorg server. -#XORG_ARGS= +XORG_ARGS=@NO_CURSOR_ARG@ @BLANK_ARGS@ # Allows to use dynamic motd script to generate custom MOTD. #DYNAMIC_MOTD=false
@@ -71,7 +71,7 @@ SESSION_ERROR_LOGGING=rotate #DEFAULT_XAUTHORITY=false #If set true, Xorg will be started as rootless, if system allows and emptty is running in daemon mode. -#ROOTLESS_XORG=false +ROOTLESS_XORG=@ROOTLESS_X_ARG@ #If set true, environmental groups are printed to differ Xorg/Wayland/Custom/UserCustom desktops. IDENTIFY_ENVS=false
meta/recipes-graphics/x11-common/xserver-nodm-init/gplv2-license.patch deleted +0 -355
355 lines not on this page · raw message
meta/recipes-graphics/x11-common/xserver-nodm-init/xserver-nodm deleted +0 -75
75 lines not on this page · raw message
meta/recipes-graphics/x11-common/xserver-nodm-init/xserver-nodm.conf.in deleted +0 -7
@@ -1,7 +0,0 @@ -# common environment file for sysvinit and systemd - -XSERVER=/usr/bin/Xorg -DISPLAY=:0 -ARGS=" -br -pn @BLANK_ARGS@ @NO_CURSOR_ARG@ " -HOME=@HOME@ -USER=@USER@
meta/recipes-graphics/x11-common/xserver-nodm-init/xserver-nodm.service.in deleted +0 -11
@@ -1,11 +0,0 @@ -[Unit] -Description=Xserver startup without a display manager - -[Service] -EnvironmentFile=/etc/default/xserver-nodm -User=@USER@ -ExecStart=/etc/xserver-nodm/Xserver -Restart=always - -[Install] -Alias=display-manager.service
meta/recipes-graphics/x11-common/xserver-nodm-init_3.0.bb +16 -41
84 lines not on this page · raw message
meta/recipes-support/user-creation/xuser-account_0.1.bb +2 -1
@@ -25,7 +25,8 @@ FILES:${PN} = "${sysconfdir}/dbus-1/system.d/system-xuser.conf" USERADD_PACKAGES = "${PN}" USERADD_PARAM:${PN} = "--create-home \ - --groups video,tty,audio,input,shutdown,disk \ + --groups video,tty,audio,input,shutdown,disk,nopasswdlogin \ --user-group xuser" +GROUPADD_PARAM:${PN} = "-r nopasswdlogin" ALLOW_EMPTY:${PN} = "1"
Signature · 1 line
2.49.0
Quoted 33 lines · rs
From: Randolph Sapp <[email protected]>
We've recently run into some issues with weston-init attempting to start Weston prior to all drm devices being registered. There's not really a good, scriptable mechanism to listen in to device registration events that works with the existing weston-init package. Well, at least one that doesn't involve polling files or introducing more dependency on the init system being used.
I also see there is also a lot of scripting around starting X11, xserver-nodm-init, that (from my limited review) should experience the same issue.
I'd like to introduce the following display manager for oe-core, emptty [1]. This display manager is, as described upstream, a "Dead simple CLI Display Manager on TTY". It supports both x11 and wayland sessions, with togglable build parameters to completely remove x11 and pam dependencies. It's licensed MIT, which shouldn't be an issue for any users. (It is written in Go, if you have opinions about that.)
With this, both weston-init and the xserver-nodm-init packages can be re-tuned to leverage this display manager and simply add a user and emptty config for an autologin session. This can resolve the current behavior across init systems without additional scripting, and move some development out of this layer.
This lists myself as a maintainer of emptty as well as xserver-nodm-init and xuser-account since these are currently unassigned and I've reworked them significantly here.
Sorry for the delay on this series. I found a few bugs in emptty that I wanted to address before submitting this officially.
Quoted 17 lines
v2:
- Address spelling issues in commit messages
- Attempt to resolve some test related issues with weston
- Add additional logs to X11 related tests
v3:
- ResetAUTOLOGIN_MAX_RETRYto the default value of 2. When running under QEMU the first auth attempt almost always fails.
v4:
- Add a tmpfile entry for the x11 domain socket directory.
- Remove some scripts associated with weston-init that were being shipped with weston
v5:
- Move tmpfile data to individual files
- Add explicit entries for these in the FILES variable
v6:
- Do not attempt to ship a tmpfiles.d entry in libx11
Hi Randolph,
Thanks for the version, but again a previously seen error. Sorry for the bad news :(
RESULTS - xorg.XorgTest.test_xorg_running: FAILED (1.31s)
...
AssertionError: 1 != 0 : Xorg does not appear to be running PID USER VSZ STAT COMMAND
https://autobuilder.yoctoproject.org/valkyrie/#/builders/20/builds/1637
https://autobuilder.yoctoproject.org/valkyrie/#/builders/95/builds/1638
https://autobuilder.yoctoproject.org/valkyrie/#/builders/9/builds/1649
I will drop the patch on my side, but I will keep the a-full build running, we might see some other failures in the coming hours:
https://autobuilder.yoctoproject.org/valkyrie/#/builders/29/builds/1630
Signature · 3 lines
Mathieu Dubois-Briand, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
Quoted 61 lines · Mathieu Dubois-Briand
Quoted 33 lines · rs
From: Randolph Sapp <[email protected]>
We've recently run into some issues with weston-init attempting to start Weston prior to all drm devices being registered. There's not really a good, scriptable mechanism to listen in to device registration events that works with the existing weston-init package. Well, at least one that doesn't involve polling files or introducing more dependency on the init system being used.
I also see there is also a lot of scripting around starting X11, xserver-nodm-init, that (from my limited review) should experience the same issue.
I'd like to introduce the following display manager for oe-core, emptty [1]. This display manager is, as described upstream, a "Dead simple CLI Display Manager on TTY". It supports both x11 and wayland sessions, with togglable build parameters to completely remove x11 and pam dependencies. It's licensed MIT, which shouldn't be an issue for any users. (It is written in Go, if you have opinions about that.)
With this, both weston-init and the xserver-nodm-init packages can be re-tuned to leverage this display manager and simply add a user and emptty config for an autologin session. This can resolve the current behavior across init systems without additional scripting, and move some development out of this layer.
This lists myself as a maintainer of emptty as well as xserver-nodm-init and xuser-account since these are currently unassigned and I've reworked them significantly here.
Sorry for the delay on this series. I found a few bugs in emptty that I wanted to address before submitting this officially.
Quoted 17 lines
v2:
- Address spelling issues in commit messages
- Attempt to resolve some test related issues with weston
- Add additional logs to X11 related tests
v3:
- ResetAUTOLOGIN_MAX_RETRYto the default value of 2. When running under QEMU the first auth attempt almost always fails.
v4:
- Add a tmpfile entry for the x11 domain socket directory.
- Remove some scripts associated with weston-init that were being shipped with weston
v5:
- Move tmpfile data to individual files
- Add explicit entries for these in the FILES variable
v6:
- Do not attempt to ship a tmpfiles.d entry in libx11Hi Randolph,
Thanks for the version, but again a previously seen error. Sorry for the bad news :(
RESULTS - xorg.XorgTest.test_xorg_running: FAILED (1.31s)
...
AssertionError: 1 != 0 : Xorg does not appear to be running PID USER VSZ STAT COMMAND
Quoted 8 lines
https://autobuilder.yoctoproject.org/valkyrie/#/builders/20/builds/1637
https://autobuilder.yoctoproject.org/valkyrie/#/builders/95/builds/1638
https://autobuilder.yoctoproject.org/valkyrie/#/builders/9/builds/1649I will drop the patch on my side, but I will keep the a-full build running, we might see some other failures in the coming hours:
https://autobuilder.yoctoproject.org/valkyrie/#/builders/29/builds/1630
Ah, the auth failures persist. Normally this is caused when the user attempts to login before the user account was created or something unusual like that. I don't suppose there's any way this could be happening on the test machines?
I'm not able to reproduce this locally, but my build config is different. I tend to use oe-core and nodistro instead of poky. They aren't playing around with the reproducible build date stuff or anything like that are they?
Well, I mean I did see this on occasion. The first login attempt almost always failed, but the subsequent attempts were fine. Bumping the attempt count back up the default (2) was enough to resolve it on my end, but evidently it's not enough here.
This is not happening on hardware from what I'm seeing.
Quoted 79 lines · Randolph Sapp
Quoted 61 lines · Mathieu Dubois-Briand
Quoted 33 lines · rs
From: Randolph Sapp <[email protected]>
We've recently run into some issues with weston-init attempting to start Weston prior to all drm devices being registered. There's not really a good, scriptable mechanism to listen in to device registration events that works with the existing weston-init package. Well, at least one that doesn't involve polling files or introducing more dependency on the init system being used.
I also see there is also a lot of scripting around starting X11, xserver-nodm-init, that (from my limited review) should experience the same issue.
I'd like to introduce the following display manager for oe-core, emptty [1]. This display manager is, as described upstream, a "Dead simple CLI Display Manager on TTY". It supports both x11 and wayland sessions, with togglable build parameters to completely remove x11 and pam dependencies. It's licensed MIT, which shouldn't be an issue for any users. (It is written in Go, if you have opinions about that.)
With this, both weston-init and the xserver-nodm-init packages can be re-tuned to leverage this display manager and simply add a user and emptty config for an autologin session. This can resolve the current behavior across init systems without additional scripting, and move some development out of this layer.
This lists myself as a maintainer of emptty as well as xserver-nodm-init and xuser-account since these are currently unassigned and I've reworked them significantly here.
Sorry for the delay on this series. I found a few bugs in emptty that I wanted to address before submitting this officially.
Quoted 17 lines
v2:
- Address spelling issues in commit messages
- Attempt to resolve some test related issues with weston
- Add additional logs to X11 related tests
v3:
- ResetAUTOLOGIN_MAX_RETRYto the default value of 2. When running under QEMU the first auth attempt almost always fails.
v4:
- Add a tmpfile entry for the x11 domain socket directory.
- Remove some scripts associated with weston-init that were being shipped with weston
v5:
- Move tmpfile data to individual files
- Add explicit entries for these in the FILES variable
v6:
- Do not attempt to ship a tmpfiles.d entry in libx11Hi Randolph,
Thanks for the version, but again a previously seen error. Sorry for the bad news :(
RESULTS - xorg.XorgTest.test_xorg_running: FAILED (1.31s)
...
AssertionError: 1 != 0 : Xorg does not appear to be running PID USER VSZ STAT COMMANDQuoted 8 lines
https://autobuilder.yoctoproject.org/valkyrie/#/builders/20/builds/1637
https://autobuilder.yoctoproject.org/valkyrie/#/builders/95/builds/1638
https://autobuilder.yoctoproject.org/valkyrie/#/builders/9/builds/1649I will drop the patch on my side, but I will keep the a-full build running, we might see some other failures in the coming hours:
https://autobuilder.yoctoproject.org/valkyrie/#/builders/29/builds/1630
Ah, the auth failures persist. Normally this is caused when the user attempts to login before the user account was created or something unusual like that. I don't suppose there's any way this could be happening on the test machines?
I'm not able to reproduce this locally, but my build config is different. I tend to use oe-core and nodistro instead of poky. They aren't playing around with the reproducible build date stuff or anything like that are they?
Quoted 6 lines
Well, I mean I did see this on occasion. The first login attempt almost always failed, but the subsequent attempts were fine. Bumping the attempt count back up the default (2) was enough to resolve it on my end, but evidently it's not enough here.
This is not happening on hardware from what I'm seeing.
Hi Randolph,
I was able to reproduce it locally, with the following configuration:
- https://git.yoctoproject.org/poky-ci-archive on tag autobuilder.yoctoproject.org/valkyrie/a-full-1630.
- template local.conf, with the following modifications:
MACHINE = "qemux86"
DISTRO = "poky-altcfg"
SDKMACHINE = "x86_64"
PACKAGE_CLASSES = "package_rpm package_deb package_ipk"
INHERIT += 'image-buildinfo'
IMAGE_BUILDINFO_VARS:append = ' IMAGE_BASENAME IMAGE_NAME'
PACKAGE_CLASSES = 'package_ipk package_rpm package_deb'
IMAGE_ROOTFS_EXTRA_SPACE:append = '${@bb.utils.contains('IMAGE_FEATURES', 'package-management', ' + 262144', '', d)}' IMAGE_INSTALL:append = ' ssh-pregen-hostkeys'
SANITY_TESTED_DISTROS = ''
OE_FRAGMENTS += 'core/yocto-autobuilder/autobuilder core/yocto-autobuilder/autobuilder-resource-constraints' # (Plus my own sstate cache and downloads dir).
- bitbake core-image-sato
- bitbake core-image-sato:do_testimage
Of course using the tag from poky-ci-archive is not mandatory, you should see the same behaviour with master branch and your patches.
I saw some oe-selftest was also failing on the autobuilder, with an error related to emptty. You can reproduce it with:
oe-selftest -r sstatetests.SStateHashSameSigs3.test_sstate_sametune_samesigs
Signature · 3 lines
Mathieu Dubois-Briand, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
Quoted 1 line · Mathieu Dubois-Briand
On Wed May 28, 2025 at 12:16 AM CEST, Randolph Sapp wrote:
[snip]
Quoted 36 lines
Quoted 8 lines
Ah, the auth failures persist. Normally this is caused when the user attempts to login before the user account was created or something unusual like that. I don't suppose there's any way this could be happening on the test machines?
I'm not able to reproduce this locally, but my build config is different. I tend to use oe-core and nodistro instead of poky. They aren't playing around with the reproducible build date stuff or anything like that are they?
Quoted 6 lines
Well, I mean I did see this on occasion. The first login attempt almost always failed, but the subsequent attempts were fine. Bumping the attempt count back up the default (2) was enough to resolve it on my end, but evidently it's not enough here.
This is not happening on hardware from what I'm seeing.
Hi Randolph,
I was able to reproduce it locally, with the following configuration:
- https://git.yoctoproject.org/poky-ci-archive on tag autobuilder.yoctoproject.org/valkyrie/a-full-1630.
- template local.conf, with the following modifications:
MACHINE = "qemux86"
DISTRO = "poky-altcfg"
SDKMACHINE = "x86_64"
PACKAGE_CLASSES= "package_rpmpackage_debpackage_ipk"
INHERIT += 'image-buildinfo'
IMAGE_BUILDINFO_VARS:append = 'IMAGE_BASENAMEIMAGE_NAME'
PACKAGE_CLASSES= 'package_ipkpackage_rpmpackage_deb'
IMAGE_ROOTFS_EXTRA_SPACE:append = '${@bb.utils.contains('IMAGE_FEATURES', 'package-management', ' + 262144', '', d)}' IMAGE_INSTALL:append = ' ssh-pregen-hostkeys'
SANITY_TESTED_DISTROS= ''
OE_FRAGMENTS+= 'core/yocto-autobuilder/autobuilder core/yocto-autobuilder/autobuilder-resource-constraints' # (Plus my own sstate cache and downloads dir).
- bitbake core-image-sato
- bitbake core-image-sato:do_testimage
Quoted 7 lines
Of course using the tag from poky-ci-archive is not mandatory, you should see the same behaviour with master branch and your patches.
I saw some oe-selftest was also failing on the autobuilder, with an error related to emptty. You can reproduce it with:
oe-selftest -r sstatetests.SStateHashSameSigs3.test_sstate_sametune_samesigs
Status report, since it's been a little bit.
Seems this some i386/x86 fault only exposed in this specific environment. I was unable to reproduce it on Archlinux32. I've traced it back to a SEGFAULT when we attempt to execute mcookie as the target user. Odd stuff. Delve and gdb aren't helping me here. Running some more tests to see if I can corner this some other way.
- Randolph
Quoted 51 lines · Randolph Sapp
Quoted 1 line · Mathieu Dubois-Briand
On Wed May 28, 2025 at 12:16 AM CEST, Randolph Sapp wrote:
[snip]
Quoted 36 lines
Quoted 8 lines
Ah, the auth failures persist. Normally this is caused when the user attempts to login before the user account was created or something unusual like that. I don't suppose there's any way this could be happening on the test machines?
I'm not able to reproduce this locally, but my build config is different. I tend to use oe-core and nodistro instead of poky. They aren't playing around with the reproducible build date stuff or anything like that are they?
Quoted 6 lines
Well, I mean I did see this on occasion. The first login attempt almost always failed, but the subsequent attempts were fine. Bumping the attempt count back up the default (2) was enough to resolve it on my end, but evidently it's not enough here.
This is not happening on hardware from what I'm seeing.
Hi Randolph,
I was able to reproduce it locally, with the following configuration:
- https://git.yoctoproject.org/poky-ci-archive on tag autobuilder.yoctoproject.org/valkyrie/a-full-1630.
- template local.conf, with the following modifications:
MACHINE = "qemux86"
DISTRO = "poky-altcfg"
SDKMACHINE = "x86_64"
PACKAGE_CLASSES= "package_rpmpackage_debpackage_ipk"
INHERIT += 'image-buildinfo'
IMAGE_BUILDINFO_VARS:append = 'IMAGE_BASENAMEIMAGE_NAME'
PACKAGE_CLASSES= 'package_ipkpackage_rpmpackage_deb'
IMAGE_ROOTFS_EXTRA_SPACE:append = '${@bb.utils.contains('IMAGE_FEATURES', 'package-management', ' + 262144', '', d)}' IMAGE_INSTALL:append = ' ssh-pregen-hostkeys'
SANITY_TESTED_DISTROS= ''
OE_FRAGMENTS+= 'core/yocto-autobuilder/autobuilder core/yocto-autobuilder/autobuilder-resource-constraints' # (Plus my own sstate cache and downloads dir).
- bitbake core-image-sato
- bitbake core-image-sato:do_testimageQuoted 7 lines
Of course using the tag from poky-ci-archive is not mandatory, you should see the same behaviour with master branch and your patches.
I saw some oe-selftest was also failing on the autobuilder, with an error related to emptty. You can reproduce it with:
oe-selftest -r sstatetests.SStateHashSameSigs3.test_sstate_sametune_samesigs
Status report, since it's been a little bit.
Quoted 7 lines
Seems this some i386/x86 fault only exposed in this specific environment. I was unable to reproduce it on Archlinux32. I've traced it back to a SEGFAULT when we attempt to execute mcookie as the target user. Odd stuff. Delve and gdb aren't helping me here. Running some more tests to see if I can corner this some other way.
- Randolph
Well, I managed to determine this is specifically an issue after loading libpam and attempting to authenticate.
Seems like loading both common-auth and common-account before calling pam_acct_mgmt results in the above described segfault.
There are 3 pam interactions made after the initial auth.
1. pam_acct_mgmt
2. pam_set_item
3. pam_setcred
Loading only common-account reports that pam_acct_mgmt fails with the PAM_SILENT flag set.
Removing that flag makes the later pam_setcred fail, but allows pam_acct_mgmt and pam_set_item to work correctly. Removing the call to pam_acct_mgmt altogether seems to allow pam_setcred to work.
Removing the call to pam_acct_mgmt also allows both common-auth and common-account to be loaded without triggering the segfault.
Very unusual. Everything points at the obvious "arch specific problems with the go pam wrapper", but other distributions don't seem to have this issue and nothing seems particularly wrong with the wrapper itself. Could be a manifestation of multiple issues though, considering how much the behavior changes depending on what pam modules are loaded.
Guess I have some reading to do this weekend.
- Randolph
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 87 lines · Randolph Sapp via lists.openembedded.org
Quoted 51 lines · Randolph Sapp
Quoted 1 line · Mathieu Dubois-Briand
On Wed May 28, 2025 at 12:16 AM CEST, Randolph Sapp wrote:
[snip]
Quoted 36 lines
Quoted 8 lines
Ah, the auth failures persist. Normally this is caused when the user attempts to login before the user account was created or something unusual like that. I don't suppose there's any way this could be happening on the test machines?
I'm not able to reproduce this locally, but my build config is different. I tend to use oe-core and nodistro instead of poky. They aren't playing around with the reproducible build date stuff or anything like that are they?
Quoted 6 lines
Well, I mean I did see this on occasion. The first login attempt almost always failed, but the subsequent attempts were fine. Bumping the attempt count back up the default (2) was enough to resolve it on my end, but evidently it's not enough here.
This is not happening on hardware from what I'm seeing.
Hi Randolph,
I was able to reproduce it locally, with the following configuration:
- https://git.yoctoproject.org/poky-ci-archive on tag autobuilder.yoctoproject.org/valkyrie/a-full-1630.
- template local.conf, with the following modifications:
MACHINE = "qemux86"
DISTRO = "poky-altcfg"
SDKMACHINE = "x86_64"
PACKAGE_CLASSES= "package_rpmpackage_debpackage_ipk"
INHERIT += 'image-buildinfo'
IMAGE_BUILDINFO_VARS:append = 'IMAGE_BASENAMEIMAGE_NAME'
PACKAGE_CLASSES= 'package_ipkpackage_rpmpackage_deb'
IMAGE_ROOTFS_EXTRA_SPACE:append = '${@bb.utils.contains('IMAGE_FEATURES', 'package-management', ' + 262144', '', d)}' IMAGE_INSTALL:append = ' ssh-pregen-hostkeys'
SANITY_TESTED_DISTROS= ''
OE_FRAGMENTS+= 'core/yocto-autobuilder/autobuilder core/yocto-autobuilder/autobuilder-resource-constraints' # (Plus my own sstate cache and downloads dir).
- bitbake core-image-sato
- bitbake core-image-sato:do_testimageQuoted 7 lines
Of course using the tag from poky-ci-archive is not mandatory, you should see the same behaviour with master branch and your patches.
I saw some oe-selftest was also failing on the autobuilder, with an error related to emptty. You can reproduce it with:
oe-selftest -r sstatetests.SStateHashSameSigs3.test_sstate_sametune_samesigs
Status report, since it's been a little bit.
Quoted 7 lines
Seems this some i386/x86 fault only exposed in this specific environment. I was unable to reproduce it on Archlinux32. I've traced it back to a SEGFAULT when we attempt to execute mcookie as the target user. Odd stuff. Delve and gdb aren't helping me here. Running some more tests to see if I can corner this some other way.
- Randolph
Well, I managed to determine this is specifically an issue after loading libpam and attempting to authenticate.
Seems like loading both common-auth and common-account before calling
pam_acct_mgmtresults in the above described segfault.There are 3 pam interactions made after the initial auth.
1.pam_acct_mgmt
2.pam_set_item
3.pam_setcredLoading only common-account reports that
pam_acct_mgmtfails with thePAM_SILENTflag set.Removing that flag makes the later
pam_setcredfail, but allowspam_acct_mgmtandpam_set_itemto work correctly. Removing the call topam_acct_mgmtaltogether seems to allowpam_setcredto work.Removing the call to
pam_acct_mgmtalso allows both common-auth and common-account to be loaded without triggering the segfault.Very unusual. Everything points at the obvious "arch specific problems with the go pam wrapper", but other distributions don't seem to have this issue and nothing seems particularly wrong with the wrapper itself. Could be a manifestation of multiple issues though, considering how much the behavior changes depending on what pam modules are loaded.
Quoted 3 lines
Guess I have some reading to do this weekend.
- Randolph
Found out that specifying the noreap option for pam_unix in the account context fixes things. Suppose go was having issues with the new signal handler being registered in _unix_run_verify_binary. Still weird that it was only really an issue on i386/x86, and only in this environment.
This means we can't inherit the common-account, we'll have to require pam_unix as part of the pam.conf directly. There's also a conditional fork in the pam_unix password path with another signal handler that we should probably guard against to prevent issues with selinux configs.
I'm still not entirely satisfied with this debug. Something seems off, but given go's use of coroutines it would make sense this would cause *some* issue. Still odd. I'll poke around a bit more.
- Randolph
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 102 lines · Randolph Sapp via lists.openembedded.org
Quoted 87 lines · Randolph Sapp via lists.openembedded.org
Quoted 51 lines · Randolph Sapp
Quoted 1 line · Mathieu Dubois-Briand
On Wed May 28, 2025 at 12:16 AM CEST, Randolph Sapp wrote:
[snip]
Quoted 36 lines
Quoted 8 lines
Ah, the auth failures persist. Normally this is caused when the user attempts to login before the user account was created or something unusual like that. I don't suppose there's any way this could be happening on the test machines?
I'm not able to reproduce this locally, but my build config is different. I tend to use oe-core and nodistro instead of poky. They aren't playing around with the reproducible build date stuff or anything like that are they?
Quoted 6 lines
Well, I mean I did see this on occasion. The first login attempt almost always failed, but the subsequent attempts were fine. Bumping the attempt count back up the default (2) was enough to resolve it on my end, but evidently it's not enough here.
This is not happening on hardware from what I'm seeing.
Hi Randolph,
I was able to reproduce it locally, with the following configuration:
- https://git.yoctoproject.org/poky-ci-archive on tag autobuilder.yoctoproject.org/valkyrie/a-full-1630.
- template local.conf, with the following modifications:
MACHINE = "qemux86"
DISTRO = "poky-altcfg"
SDKMACHINE = "x86_64"
PACKAGE_CLASSES= "package_rpmpackage_debpackage_ipk"
INHERIT += 'image-buildinfo'
IMAGE_BUILDINFO_VARS:append = 'IMAGE_BASENAMEIMAGE_NAME'
PACKAGE_CLASSES= 'package_ipkpackage_rpmpackage_deb'
IMAGE_ROOTFS_EXTRA_SPACE:append = '${@bb.utils.contains('IMAGE_FEATURES', 'package-management', ' + 262144', '', d)}' IMAGE_INSTALL:append = ' ssh-pregen-hostkeys'
SANITY_TESTED_DISTROS= ''
OE_FRAGMENTS+= 'core/yocto-autobuilder/autobuilder core/yocto-autobuilder/autobuilder-resource-constraints' # (Plus my own sstate cache and downloads dir).
- bitbake core-image-sato
- bitbake core-image-sato:do_testimageQuoted 7 lines
Of course using the tag from poky-ci-archive is not mandatory, you should see the same behaviour with master branch and your patches.
I saw some oe-selftest was also failing on the autobuilder, with an error related to emptty. You can reproduce it with:
oe-selftest -r sstatetests.SStateHashSameSigs3.test_sstate_sametune_samesigs
Status report, since it's been a little bit.
Quoted 7 lines
Seems this some i386/x86 fault only exposed in this specific environment. I was unable to reproduce it on Archlinux32. I've traced it back to a SEGFAULT when we attempt to execute mcookie as the target user. Odd stuff. Delve and gdb aren't helping me here. Running some more tests to see if I can corner this some other way.
- Randolph
Well, I managed to determine this is specifically an issue after loading libpam and attempting to authenticate.
Seems like loading both common-auth and common-account before calling
pam_acct_mgmtresults in the above described segfault.There are 3 pam interactions made after the initial auth.
1.pam_acct_mgmt
2.pam_set_item
3.pam_setcredLoading only common-account reports that
pam_acct_mgmtfails with thePAM_SILENTflag set.Removing that flag makes the later
pam_setcredfail, but allowspam_acct_mgmtandpam_set_itemto work correctly. Removing the call topam_acct_mgmtaltogether seems to allowpam_setcredto work.Removing the call to
pam_acct_mgmtalso allows both common-auth and common-account to be loaded without triggering the segfault.Very unusual. Everything points at the obvious "arch specific problems with the go pam wrapper", but other distributions don't seem to have this issue and nothing seems particularly wrong with the wrapper itself. Could be a manifestation of multiple issues though, considering how much the behavior changes depending on what pam modules are loaded.
Quoted 3 lines
Guess I have some reading to do this weekend.
- Randolph
Found out that specifying the noreap option for
pam_unixin the account context fixes things. Suppose go was having issues with the new signal handler being registered in_unix_run_verify_binary. Still weird that it was only really an issue on i386/x86, and only in this environment.This means we can't inherit the common-account, we'll have to require
pam_unixas part of the pam.conf directly. There's also a conditional fork in thepam_unixpassword path with another signal handler that we should probably guard against to prevent issues with selinux configs.
Quoted 5 lines
I'm still not entirely satisfied with this debug. Something seems off, but given go's use of coroutines it would make sense this would cause *some* issue. Still odd. I'll poke around a bit more.
- Randolph
Bit of a status update. Sorry for the delay, this got the backburner for far too long. I isolated this to specifically the operation of switching the signal handler back to the go runtime one.
As to why this was only viewable under yocto, it seems that yocto is the only env that feels the need to run the binary helper for shadowed users. Not really sure why, but that's more configuration prone and not relevant to the more glaring issue.
Sigaction when used in libraries linked to cgo applications can misbehave as it uses the glibc version of sigaction. Cgo applications actually use a go definition of sigaction with some additional checks depending on the OS and architecture selected. Seems like only i386/x86 platforms were actually misbehaving, but it's something to note moving forward.
There are upstream tickets to track the current debug:
https://github.com/msteinert/pam/issues/33
https://github.com/golang/go/issues/75253
I'll refrain from updating this thread until those are closed.
- Randolph