FOSS activity in July 2024
- I continued participating in Debian kernel team meetings.
- For the Debian linux package:
- (LTS) I investigated a regression for nftables introduced in my final upload of linux to buster-security, and passed on the information to the Freexian ELTS team.
- I uploaded:
- linux version 6.1.94-1~bpo11+1 to bullseye-backports.
- linux version 6.8.12-1~bpo12+1 to bookworm-backports.
- linux version 6.9.7-1~bpo12+1 to bookworm-backports.
- linux version 6.10-1~exp1 to experimental.
- linux version 6.1.99-1~bpo11+1 to bullseye-backports (but it was never accepted).
- linux version 6.10.1-1~exp1 to experimental.
- linux version 6.9.10-1~bpo12+1 to bookworm-backports.
- I opened or updated MRs:
- !1077: d/b/gencontrol.py, d/rules.real: Restore config checks on kernels to be signed
- !1112: Update d/l/p/debian_linux/firmware.py for current WHENCE format
- !1115: Update to 6.10-rc7
- !1119: Update d/b/test-patches to work with current package
- !1126: [alpha] scsi: Disable SCSI_IMM (fixes FTBFS)
- !1133: Draft: Fix sh4/sh7785lcr flavour
- I reviewed MRs:
- !675: [arm64] drivers/usb/host: Enable USB_XHCI_PCI_RENESAS as module (Closes: #1032671)
- !732: [x86] linux-cpupower: Add intel-speed-select command
- !957: debian/bin/gencontrol.py: allow adding a custom suffix to the abiname (closed)
- !964: tools/arch/x86/intel_sdsi: Add sdsi package for Intel SDSi provisioning tool
- !1037: debian/rules.real: set absolute bpftool path for linux 6.8+ (closed)
- !1038: debian/rules.real: export LANG = C.UTF-8 for sphinx
- !1041: Add “-b” flag to genorig.py
- !1051: [x86] drivers/platform/x86: Enable MSI_EC as module (merged)
- !1059: [amd64/cloud] drivers/watchdog: Enable I6300ESB_WDT as module (merged)
- !1074: MIPS64EL: add mips64r6el flavor (merged)
- !1084: Remove unused check for image size
- !1093: d/rules.d/t/perf/Makefile: Enable debuginfod support. (merged)
- !1094: [arm64] drivers/gpu/drm/bridge/synopsys: Enable DRM_DW_HDMI_I2S_AUDIO as module (merged)
- !1095: [arm64] Enable config options for Qualcomm boards (merged)
- !1100: kernel/power: enable CONFIG_HIBERNATION_COMP_LZ4
- !1118: [x86] sound/soc/intel/avs/boards: Enable SND_SOC_INTEL_AVS_MACH_MAX98927 as a module (merged)
- !1122: Enable snd_soc_pcm5102a as a module (merged)
- !1123: [ppc64*] Switch default kernel to 4k page size (merged)
- !1128: drivers/md/dm-vdo: Enable DM_VDO as module (merged)
- !1129: Backport Microsoft Azure Network Adapter from 6.10
- !1134: debian/rules: sort control.md5sums to improve reproducibility (merged)
- !1135: [arm64] Re-enable RELR (merged)
- !1136: Compile with gcc-14 on all architectures
- !1139: [arm64] enable CONFIG_QCOM_LMH, another SDM845-related option (merged)
- !1141: drivers/net: Enable NETKIT (BPF-programmable network device)
- !1142: fs/erofs: Enable more EROFS compression algorithms (merged)
- I merged my own MRs:
- !1110: d/l/p/debian_linux/firmware.py: Handle RawFile fields
- !1112: Update d/l/p/debian_linux/firmware.py for current WHENCE format
- !1119: Update d/b/test-patches to work with current package
- !1126: [alpha] scsi: Disable SCSI_IMM (fixes FTBFS)
- (LTS) To support Debian ELTS, I created branches of the Linux 5.10 and 6.1 packaging with backports of the change to use an ephemeral module signing key.
- I answered a query about use of the linux-image-*-unsigned packages.
- I responded to bug reports:
- #989229: grub-install: warning: Cannot read EFI Boot* variables
- #1039883: linux: ext4 corruption with symlinks
- #1063754: fat-modules: SD corruption upon opening file on Linux desktop
- #1075855: Kernel panic caused by aacraid module prevents normal boot
- #1072063: one of the external monitors randomly blank for 2-3 seconds with 6.8/6.9 Linux kernels (regression)
- #1072311: linux-perf can (and should) link against libdebuginfod
- Upstream, I commented on how to detect 32-bit architectures in order to fix CVE-2024-42258.
- Upstream, I submitted the patch xhci-pci: Make xhci-pci-renesas a proper modular driver which is a prerequisite for merging MR !675.
- I asked the Debian Super-H porters whether the sh7785lcr kernel flavour was useful.
- In dput-ng, I merged my own MR !36: rsync, scp: Fix username lookup.
- In devscripts, I updated and merged my own MR !292: uscan: Allow compression of VCS exports to be disabled. This can make uscan a lot faster for packages that use a VCS as upstream and exclude some files from it.
- For the Debian firmware-nonfree package:
- I opened MRs:
- I reviewed MRs:
- I merged my own MRs:
- !96: Update to 20240610
- !98: Include or exclude most unpackaged firmware
- !101: Update to 20240709 and remove some file exclusions
- I uploaded versions 20240610-1 and 20240709-1 to unstable.
- I responded to bug reports:
- In the kernel-team repository:
- I reviewed MRs:
- I deleted the obsolete script that !2 would have updated.
- For the Debian wireless-regdb package:
- I reviewed MRs:
- (LTS) !4: merge stretch-elts 2022.04.08-1~deb9u1 upload (closed)
- (LTS) !5: Upload For LTS (buster) (merged)
- I reviewed MRs:
- For the Debian nfs-utils package:
- I opened MR !31: Fixes for handling of state files in /var/lib/nfs in response to bug #1074359: nfs-kernel-server: Updating package unexports all filesystems, and later merged it.
- I reviewed and merged MR !15: A couple more DEP8 tests.
- For the Debian klibc package:
- For the Debian ktls-utils package:
- I updated to upstream version 0.11 and uploaded version 0.11-1 to unstable.
- For the Debian initramfs-tools package:
- I uploaded version 0.143.1 to unstable, with no changes from version 0.143. One of the changes in 0.143 happened to fix the newly reported #1076539: plymouth: Updating plymouth fails with “No space left on device” (and its many duplicates).
- I reviewed MRs:
- !70: Support MODULES=dep usage when root was mounted from root specified on kernel command line (closed)
- !78: feature: safely close devices on shutdown (closed)
- !84: Allow providing UDEV_WAIT and ROUNDTTT times in environment variables
- !89: init: Remove tmpfs from rootfstype option
- !96: mkinitramfs: Do not store intermediate main cpio archive (merged)
- !107: Replace copy_modules_dir by manual_add_modules calls (merged)
- !116: autopkgtest: Enable KVM if available (merged)
- !117: install hid-multitouch module for Surface Pro 4 Keyboard (merged)
- !118: fsck: Mention file system name in failed identification warning (merged)
- !119: Fix resume device type check
- !120: hook-functions: auto_add_modules: Add onboard_usb_hub, onboard_usb_dev (merged)
- !121: hook-functions: add_loaded_modules: Walk bound devices for suppliers (merged)
- !122: d/gbp.conf: Set gbp-dch options matching existing changelog entries (merged)
- !123: mkinitramfs: Add -m argument to override MODULES setting (merged)
- !124: mkinitramfs: Add MODULES=all option to add every module (closed)
- !126: Move shellcheck configuration to .shellcheckrc (merged)
- I responded to bug reports:
- #961395: initramfs-tools: failed hardlink initrd.img
- #980021: initramfs-tools: Upgrading a LVM2 system with separate /usr to buster breaks booting
- #1027749: update-initramfs could diagnose attempt to run with /dev not mounted
- #1054991: initramfs-tools: failed to make backup on esp directory /boot
- #1065698: update-initramfs: -k all stopped working
- #1068195: USB keyboard unusable when booting with init=/bin/bash
- I reported Debian bugs:
- For the Debian a56 package, which is a build-dependency of firmware-free, I made an NMU fixing a build failure with gcc-14 and many compiler warnings. These changes were included in version 1.3+dfsg-11.