Zander671/kdumpst
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Multiple fixes and improvements

Browse Source 
OK, likely I should have split this commit in many more, to ease
reviews and improve bisectability. But since I didn't, I'll at least
try to summarize here what was changed by this big commit, starting
with the more complex additions to the trivial ones:

- Added a log submission system; for now, it doesn't submit logs
to anywhere since we don't have such API well-defined. But it saves
the logs locally in a tar.zst, using timestamps to define the moment
of log collection and clears the old files. The naming of the log
compressed tarball makes use of the Deck serial number and Steam
most recent logged account (thanks @TonyP for that idea).

- Still on the log submission topic: we try first pstore logs, and
if we have none of them, we check kdump logs. In case we have a
vmcore, we save it locally (after renaming), but for now we consider
that this kind of file won't be submitted.

- Since we now have makedumpfile in Holo (thanks @xexaxo), I've
removed this binary from here and made this package dependent on
makedumpfile. Also, adjusted the respective scripts.

- Added some error messages through logger, in case we fail to
load pstore/kdump or fail in the log submission script.

- Changed systemd service to just do the pstore/kdump loading
and call the submit_report.sh script, which is then "disowned"
in order we finish as fast as possible the systemd service; boot
delays due to this service wouldn't be nice.

- Increased the "crashkernel" memory recommended in the documentation,
since I've noticed the most recent version of the kernel requires
a bit more - let's play safe!

- Changed timestamps to use UTC tz - this is due kdump collection
happening so early, that timezone is not set, so let's stick with
UTC in all cases.

- Checked scripts with shellcheck[0] and improved the README.

[0] Accepted most suggestions, but some are polemic, and may introduce
issues, so the scripts are not fully passing shellcheck and I don't
expect them to be in the future, it's just a minor style improvement
and failsafe for multiple shells.

Signed-off-by: Guilherme G. Piccoli <gpiccoli@igalia.com>
This commit is contained in:
Guilherme G. Piccoli
2022-01-12 18:38:32 -03:00
parent 2ab4eae53e
commit 3638dc8264
10 changed files with 189 additions and 71 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
PKGBUILD
View File

@ -4,10 +4,10 @@
# Maintainer: Guilherme G. Piccoli <gpiccoli@igalia.com>
pkgname=kdump-steamos
pkgver=0.2
pkgver=0.3
pkgrel=1
pkgdesc="Kdump scripts to collect vmcore/dmesg in a small dracut-based initramfs"
depends=('dracut' 'kexec-tools' 'systemd' 'zstd')
depends=('dmidecode' 'dracut' 'kexec-tools' 'makedumpfile' 'systemd' 'zstd')
arch=('x86_64')
license=('GPL2')
install=kdump-steamos.install
@ -17,20 +17,20 @@ source=('kdump_collect.sh'
'kdump_load.sh'
'kdump-steamos.install'
'kdump-steamos.service'
'makedumpfile'
'module-setup.sh'
'README.md'
'submit_report.sh')
'submit_report.sh'
'submitter_load.sh')
sha256sums=('38a3636c95cb97b33a71cfb2b95ccbf7a9a565e86b2128299ea7844d1135fe07'
'38751d1fa1607fc99607423a0051a2b3322db5579906401b40c11c10edd6bbc6'
'888024a0b121102688d0384cf00dca06d55d3c2fc6b18a3de0da1fc8b5c10066'
sha256sums=('2514f79a496f76af847e262eadd55a5c2f8d95375cc513efa8cadd4cd98fe1d2'
'd0ac5e7e38fa1d3355eacdf70188483456f53d3e2b18cd161dea3df87b0b8f9c'
'8a556a9ebbda88dfd29b9620a0f2e7dbea19cd5fc019eb5dc4ebf7c80e4bf238'
'06b38bd9f09da5fb22a765b6f1945fc349cc5f9d13cd32c9218b9b60b40a9010'
'6063ed2283743d8d84a89d9f3c950e5f50adf99bba5ce865a25282081ebc04c2'
'86ef2bd71551598f392fe278507449c1c872e0d42b27600cfeb5bcf9a75aa881'
'c3ceaf77021e49c3ec884e3959f49b0cbf5e8e89ad3f17d485d895d9e91725f4'
'01432491df80dfd37c6f261c17f55c574e8898003642334a4d61f8d93aef08c3'
'956efe1589d8d6533a231d8bdec6ac5cd4c1d1494b1f44b8494fe1d75f6a1e4e')
'12a9124b907f208471ba7aaac0f3261cbbd34a168cce3260fa9e7793994beebd'
'26bc2b64af0d468f050c0e0dd9e2053176d56886edad9146bc495797bf2c5810'
'b87fb8e4c4602f8ddc3b0bf6d6175d0ee7b9e0942f4dca8f1b958ed3ad445470'
'd8b432dc1602e330e61c91a8f9e6761273ff2ca8129f457828ff0d20ac6d5b25'
'cbb207ecc0f6bacefbeed41f0d4910daac6500ac2345366e1f95f09a7653c65a')
package() {
install -D -m0644 kdump.etc "$pkgdir/etc/default/kdump"
@ -42,6 +42,6 @@ package() {
install -D -m0644 README.md "$pkgdir/usr/lib/dracut/modules.d/55kdump/README"
install -D -m0755 kdump_load.sh "$pkgdir/usr/lib/kdump/kdump_load.sh"
install -D -m0755 makedumpfile "$pkgdir/usr/lib/kdump/makedumpfile"
install -D -m0755 submit_report.sh "$pkgdir/usr/lib/kdump/submit_report.sh"
install -D -m0755 submitter_load.sh "$pkgdir/usr/lib/kdump/submitter_load.sh"
}

24
README.md
View File

@ -15,7 +15,8 @@
# collection, that only grabs dmesg, and a more complete setting to grab the
# whole (compressed) vmcore. The tunnings are available at /etc/default/kdump.
#
# Also, the infrastructure is able to configure and save pstore-RAM logs.
# Also, the infrastructure is able to configure and save pstore-RAM logs;
# this is the default option.
#
# After installation and a reboot, things should be all set EXCEPT for GRUB
# config - please check the CAVEATS/INSTRUCTIONS section below. Notice the
@ -27,17 +28,17 @@
# CAVEATS / INSTRUCTIONS
# ###########################################################################
# (a) For now, we don't automatically edit any GRUB config, so the minimum
# necessary action after installing this package is to add "crashkernel=160M"
# necessary action after installing this package is to add "crashkernel=192M"
# to your GRUB config in order subsequent boots pick this setting and do reserve
# the memory, or else kdump cannot work. The memory amount was empirically
# determined - 128M wasn't enough and 144M is unstable, so 160M seems good enough.
# determined - 144M wasn't enough and 160M is unstable, so 192M seems good enough.
# If you prefer to rely on pstore-RAM, no GRUB setting should be required; this
# is currently the default (see /etc/default/kdump).
#
# (b) It requires (obviously) a RW rootfs - we've used tune2fs in order to make
# it read-write, since it's RO by default. Also, we assume the nvme partition
# scheme is default across all versions and didn't change with new updates
# for example - kdump relies in mounting partitions, etc.
# for example - both kdump and pstore relies in mounting partitions, etc.
#
# (c) Due to a post-transaction hook executed by libalpm (90-dracut-install.hook),
# unfortunately after installing the kdump-steamos package *all* initramfs images
@ -45,9 +46,9 @@
# but for now be prepared: the installation take some (long) minutes due to that ={
#
# (d) Unfortunately makedumpfile from Arch Linux is not available on official
# repos, only in AUR. So, we're hereby _packing the binary_ with all the scripts,
# which is a temporary workaround and should be resolved later - already started
# to "lobby" for package inclusion in the official channels:
# repos, only in AUR. But it is available on Holo, so we make use of that.
# Also, a discussion was started to get it included on official repos:
# https://lists.archlinux.org/pipermail/aur-general/2022-January/036767.html
# https://aur.archlinux.org/packages/makedumpfile/#comment-843853
#
#
@ -68,13 +69,16 @@
# in the past and relying in sysrq reboot as a quirk managed to be a safe option,
# so this is something to think about here. Should be easy to implement.
#
# (5) Maybe a good idea would be to allow creating the minimum image for any
# specified kernel, not only for the running one (which is what we do now).
# Low-priority idea, easy to implement.
# (5) The log submission mechanism is incomplete - we save the logs as tar.zst
# files, but they are not submitted to any remote server, etc.
#
# (6) Pstore ramoops backend has some limitations that we're discussing with
# the kernel community - right now we can only collect ONE dmesg and its
# size is truncated on "record_size" bytes, not allowing a file split like
# efi-pstore; hopefully we can improve that.
#
# (7) Maybe a good idea would be to allow creating the minimum image for any
# specified kernel, not only for the running one (which is what we do now).
# Low-priority idea, easy to implement.
#
```

2
kdump-steamos.service
View File

@ -10,7 +10,7 @@ Description=SteamOS kdump loader boot-time service
Type=oneshot
StandardOutput=journal
ExecStartPre=/usr/lib/kdump/kdump_load.sh
ExecStart=/usr/lib/kdump/submit_report.sh
ExecStart=/usr/lib/kdump/submitter_load.sh
RemainAfterExit=yes
[Install]

5
kdump.etc
View File

@ -33,3 +33,8 @@ MAKEDUMPFILE_DMESG_CMD="--dump-dmesg"
# relies in having an available RAM buffer on /proc/iomem with at least 5MiB
# in size.
USE_PSTORE_RAM=1
# This is a log submission setting, based on Steam config files, and
# *should not* be changed, or else the log sending mechanism will be
# impaired.
LOGINVDF="/home/doorstop/.local/share/Steam/config/loginusers.vdf"

8
kdump_collect.sh
View File

@ -14,7 +14,7 @@
. /usr/lib/kdump/kdump.etc
VMCORE="/proc/vmcore"
KDUMP_TIMESTAMP=$(date +"%Y%m%d%H%M")
KDUMP_TIMESTAMP=$(date -u +"%Y%m%d%H%M")
KDUMP_FOLDER="/kdump_path/${KDUMP_FOLDER}/crash/${KDUMP_TIMESTAMP}"
# Bail out in case we don't have a vmcore, i.e. either we're not kdumping
@ -31,11 +31,11 @@ fi
mkdir -p "${KDUMP_FOLDER}"
/usr/lib/kdump/makedumpfile ${MAKEDUMPFILE_DMESG_CMD} $VMCORE "${KDUMP_FOLDER}/dmesg.txt"
/usr/bin/makedumpfile ${MAKEDUMPFILE_DMESG_CMD} $VMCORE "${KDUMP_FOLDER}/dmesg.txt"
sync "${KDUMP_FOLDER}/dmesg.txt"
if [ ${FULL_COREDUMP} -ne 0 ]; then
/usr/lib/kdump/makedumpfile ${MAKEDUMPFILE_COREDUMP_CMD} $VMCORE "${KDUMP_FOLDER}/vmcore.compressed"
if [ "${FULL_COREDUMP}" -ne 0 ]; then
/usr/bin/makedumpfile ${MAKEDUMPFILE_COREDUMP_CMD} $VMCORE "${KDUMP_FOLDER}/vmcore.compressed"
sync "${KDUMP_FOLDER}/vmcore.compressed"
fi

28
kdump_load.sh
View File

@ -1,4 +1,4 @@
#!/bin/sh
#!/bin/bash
#
# SPDX-License-Identifier: LGPL-2.1+
#
@ -21,22 +21,22 @@ fi
DEVN_MOUNTED=$(mount |grep "${MOUNT_DEVNODE}" | head -n1 | cut -f3 -d\ )
KDUMP_FOLDER="${DEVN_MOUNTED}/${KDUMP_FOLDER}"
echo "${KDUMP_FOLDER}" > ${KDUMP_MNT}
sync ${KDUMP_MNT}
echo "${KDUMP_FOLDER}" > "${KDUMP_MNT}"
sync "${KDUMP_MNT}"
if [ "$1" == "initrd" ]; then
if [ "$1" = "initrd" ]; then
mkdir -p "${KDUMP_FOLDER}"
rm -f "${KDUMP_FOLDER}/kdump-initrd-$(uname -r).img"
echo "Creating the kdump initramfs for kernel \"$(uname -r)\" ..."
dracut --no-early-microcode --host-only -q -m\
"bash systemd systemd-initrd systemd-sysusers modsign dbus-daemon kdump dbus udev-rules dracut-systemd base fs-lib shutdown"\
--kver $(uname -r) "${KDUMP_FOLDER}/kdump-initrd-$(uname -r).img"
--kver "$(uname -r)" "${KDUMP_FOLDER}/kdump-initrd-$(uname -r).img"
exit 0
fi
if [ "$1" == "clear" ]; then
if [ "$1" = "clear" ]; then
rm -f ${KDUMP_FOLDER}/kdump-initrd-*
exit 0
fi
@ -46,29 +46,37 @@ fi
# here a 5MiB memory region.
# Notice that we assume ramoops is a module here - if built-in, we should
# properly load it through command-line parameters.
if [ ${USE_PSTORE_RAM} -eq 1 ]; then
if [ "${USE_PSTORE_RAM}" -eq 1 ]; then
MEM_REQUIRED=5242880 # 5MiB
RECORD_SIZE=0x200000 # 2MiB
RANGE=$(grep "RAM buffer" /proc/iomem | head -n1 | cut -f1 -d\ )
MEM_END=$(echo $RANGE | cut -f2 -d\-)
MEM_START=$(echo $RANGE | cut -f1 -d\-)
MEM_END=$(echo "$RANGE" | cut -f2 -d\-)
MEM_START=$(echo "$RANGE" | cut -f1 -d\-)
MEM_SIZE=$(( 16#${MEM_END} - 16#${MEM_START} ))
if [ ${MEM_SIZE} -ge ${MEM_REQUIRED} ]; then
if modprobe ramoops mem_address=0x${MEM_START} mem_size=${MEM_REQUIRED} record_size=${RECORD_SIZE}; then
exit 0
fi
logger "pstore-RAM load was attempted and failed...will try kdump"
fi
# Fallbacks to kdump load - if we fail when configuring pstore, better try kdump;
# who knows and we may be lucky enough to have some crashkernel reserved memory...
# TODO (maybe): could invert the order and try kdump first, if it fails, try pstore!
fi
# TODO: insert code here to validate that crashkernel is configured and
# memory is reserved; if not, set it on grub.cfg and recreate the EFI grub
# config file, warning users that in the current boot kdump is not set.
# Stolen from Debian kdump
KDUMP_CMDLINE=$(sed -re 's/(^| )(crashkernel|hugepages|hugepagesz)=[^ ]*//g;s/"/\\\\"/' /proc/cmdline)
KDUMP_CMDLINE="${KDUMP_CMDLINE} panic=-1 oops=panic fsck.mode=force fsck.repair=yes nr_cpus=1 reset_devices"
VMLINUX="$(grep -o 'BOOT_IMAGE=[^ ]*' /proc/cmdline)"
kexec -s -p "${VMLINUX#*BOOT_IMAGE=}" --initrd "${KDUMP_FOLDER}/kdump-initrd-$(uname -r).img" --append="${KDUMP_CMDLINE}" || true
if ! kexec -s -p "${VMLINUX#*BOOT_IMAGE=}" --initrd "${KDUMP_FOLDER}/kdump-initrd-$(uname -r).img" --append="${KDUMP_CMDLINE}"; then
logger "kdump load was attempted and failed"
exit 0
fi

BIN
makedumpfile
View File

Binary file not shown.

1
module-setup.sh
View File

@ -28,6 +28,7 @@ install() {
# Install necessary binaries
inst date
inst sync
inst makedumpfile
mkdir -p $initdir/usr/lib/kdump
cp -LR --preserve=all /usr/lib/kdump/* $initdir/usr/lib/kdump/

152
submit_report.sh
View File

@ -4,50 +4,138 @@
#
# Copyright (c) 2021 Valve.
#
# This is currently a dummy script for kdump, but collects and clears
# pstore saved logs for now. It aims, in the future, to submit error/crash
# reports to Valve servers in order to do a post-mortem analysis.
# This is part of SteamOS kdump - it is invoked by systemd on boot time
# and should always exit graciously to avoid breaking the boot services.
# This is the SteamOS kdump/pstore log collector and submitter; this script
# prepares the pstore/kdump collected data and submit it to the services that
# handle support at Valve. It considers pstore as a first alternative, if no
# logs found (or if pstore is not mounted for some reason), tries to check
# if kdump logs are present.
#
# We do some validation to be sure KDUMP_MNT pointed path is valid...
# That and having a valid /etc/default/kdump are essential conditions.
if [ ! -f "/etc/default/kdump" ]; then
logger "/etc/default/kdump not present - aborting..." || true
exit 0
fi
. /etc/default/kdump
# Yeah, we assume pstore is mounted by default, in this location;
KDUMP_MAIN_FOLDER="$(cat "${KDUMP_MNT}")"
rm -f "${KDUMP_MNT}"
if [ ! -d "${KDUMP_MAIN_FOLDER}" ]; then
logger "invalid folder: ${KDUMP_MAIN_FOLDER} - aborting..." || true
exit 0
fi
LOGS_FOUND=0
KDUMP_LOGS_FOLDER="${KDUMP_MAIN_FOLDER}/logs"
# Use UTC timezone to match kdump collection
CURRENT_TSTAMP=$(date -u +"%Y%m%d%H%M")
# We assume pstore is mounted by default, in this location;
# if not, we get a 0 and don't loop.
PSTORE_CNT=$(find /sys/fs/pstore/* 2>/dev/null | grep -c ramoops)
if [ ${PSTORE_CNT} -eq 0 ]; then
exit 0
fi
if [ "${PSTORE_CNT}" -ne 0 ]; then
# We do some validation to be sure KDUMP_MNT pointed path is valid...
KDUMP_CRASH_FOLDER="$(cat ${KDUMP_MNT})"
rm -f ${KDUMP_MNT}
# Dump the pstore logs in the <...>/kdump/logs/pstore subfolder.
PSTORE_FOLDER="${KDUMP_LOGS_FOLDER}/pstore"
mkdir -p "${PSTORE_FOLDER}"
if [ ! -d "${KDUMP_CRASH_FOLDER}" ]; then
exit 0
fi
# If valid, then dump the pstore logs in the crash subfolder.
KDUMP_CRASH_FOLDER="${KDUMP_CRASH_FOLDER}/crash"
mkdir -p ${KDUMP_CRASH_FOLDER}
PSTORE_TSTAMP=$(date +"%Y%m%d%H%M")
LOOP_CNT=0
while [ ${PSTORE_CNT} -gt 0 ];
do
LOOP_CNT=0
while [ "${PSTORE_CNT}" -gt 0 ]; do
PSTORE_FILE="$(find /sys/fs/pstore/* | grep ramoops | sort | head -n1)"
SAVED_FILE="${KDUMP_CRASH_FOLDER}/dmesg-pstore.${PSTORE_TSTAMP}-${LOOP_CNT}"
SAVED_FILE="${PSTORE_FOLDER}/dmesg-pstore.${CURRENT_TSTAMP}-${LOOP_CNT}"
cat ${PSTORE_FILE} > ${SAVED_FILE}
sync ${SAVED_FILE}
rm -f ${PSTORE_FILE}
cat "${PSTORE_FILE}" > "${SAVED_FILE}"
sync "${SAVED_FILE}"
rm -f "${PSTORE_FILE}"
PSTORE_CNT=$((${PSTORE_CNT} - 1))
LOOP_CNT=$((${LOOP_CNT} + 1))
done
exit 0
PSTORE_CNT=$((PSTORE_CNT - 1))
LOOP_CNT=$((LOOP_CNT + 1))
done
LOGS_FOUND=${LOOP_CNT}
# Enter the else block in case we don't have pstore logs - maybe we
# have kdump logs then.
else
KDUMP_CRASH_FOLDER="${KDUMP_MAIN_FOLDER}/crash"
KDUMP_CNT=$(find ${KDUMP_CRASH_FOLDER}/* -type d 2>/dev/null | wc -l)
if [ "${KDUMP_CNT}" -ne 0 ]; then
# Dump the kdump logs in the <...>/kdump/logs/kdump subfolder.
KD_FOLDER="${KDUMP_LOGS_FOLDER}/kdump"
mkdir -p "${KD_FOLDER}"
LOOP_CNT=0
while [ "${KDUMP_CNT}" -gt 0 ]; do
CRASH_CURRENT=$(find ${KDUMP_CRASH_FOLDER}/* -type d 2>/dev/null | head -n1)
CRASH_TSTAMP=$(basename "${CRASH_CURRENT}")
if [ -s "${CRASH_CURRENT}/dmesg.txt" ]; then
SAVED_FILE="${KD_FOLDER}/dmesg-kdump.${CRASH_TSTAMP}"
mv "${CRASH_CURRENT}/dmesg.txt" "${SAVED_FILE}"
sync "${SAVED_FILE}"
fi
# We don't care about submitting a vmcore, but let's save it if such file exists.
if [ -s "${CRASH_CURRENT}/vmcore.compressed" ]; then
SAVED_FILE="${KDUMP_CRASH_FOLDER}/vmcore.${CRASH_TSTAMP}"
mv "${CRASH_CURRENT}/vmcore.compressed" "${SAVED_FILE}"
sync "${SAVED_FILE}"
fi
rm -rf "${CRASH_CURRENT}"
KDUMP_CNT=$((KDUMP_CNT - 1))
LOOP_CNT=$((LOOP_CNT + 1))
done
LOGS_FOUND=$((LOGS_FOUND + LOOP_CNT))
fi
fi
# If we have pstore and/or kdump logs, let's process them in order to submit...
if [ ${LOGS_FOUND} -ne 0 ]; then
PNAME="$(dmidecode -s system-product-name)"
if [ "${PNAME}" = "Jupiter" ]; then
SN="$(dmidecode -s system-serial-number)"
else
SN=0
fi
STEAM_ACCOUNT=0
if [ -s "${LOGINVDF}" ]; then
# The following awk command was borrowed from:
# https://unix.stackexchange.com/a/663959
NUMREG=$(grep -c AccountName "${LOGINVDF}")
IDX=1
while [ ${IDX} -le "${NUMREG}" ]; do
MR=$(awk -v n=${IDX} -v RS='}' 'NR==n{gsub(/.*\{\n|\n$/,""); print}' "${LOGINVDF}" | grep "MostRecent" | cut -f4 -d\")
if [ "$MR" -ne 1 ]; then
IDX=$((IDX + 1))
continue
fi
STEAM_ACCOUNT=$(awk -v n=${IDX} -v RS='}' 'NR==n{gsub(/.*\{\n|\n$/,""); print}' "${LOGINVDF}" | grep "AccountName" | cut -f4 -d\")
break
IDX=$((IDX + 1))
done
fi
LOG_FNAME="steamos-${SN}-${STEAM_ACCOUNT}.${CURRENT_TSTAMP}.tar"
LOG_FNAME="${KDUMP_MAIN_FOLDER}/${LOG_FNAME}"
tar cf "${LOG_FNAME}" "${KDUMP_LOGS_FOLDER}" 1>/dev/null 2>&1
zstd < "${LOG_FNAME}" > "${LOG_FNAME}.zst"
sync "${LOG_FNAME}.zst"
rm -rf "${KDUMP_LOGS_FOLDER}" "${LOG_FNAME}"
# TODO: implement a log submission mechanism, in order to send the zstd file
# to Valve servers through an API.
fi

12
submitter_load.sh Normal file
View File

@ -0,0 +1,12 @@
#!/bin/sh
#
# SPDX-License-Identifier: LGPL-2.1+
#
# Copyright (c) 2022 Valve.
#
# This is the systemd loader for the SteamOS kdump log submitter;
# it's invoked by systemd, basically it just loads a detached
# process and exits successfuly, in order to prevent boot hangs.
/usr/lib/kdump/submit_report.sh & disown
exit 0