On 3/12/23 13:11, Sourabh Jain wrote:
The Problem:
============
Post hotplug/DLPAR events the capture kernel holds stale information about the
system. Dump collection with stale capture kernel might end up in dump capture
failure or an inaccurate dump collection.
Existing solution:
==================
The existing solution to keep the capture kernel up-to-date by monitoring
hotplug event via udev rule and trigger a full capture kernel reload for
every hotplug event.
Shortcomings:
------------------------------------------------
- Leaves a window where kernel crash might not lead to a successful dump
collection.
- Reloading all kexec components for each hotplug is inefficient.
- udev rules are prone to races if hotplug events are frequent.
More about issues with an existing solution is posted here:
- https://lkml.org/lkml/2020/12/14/532
- https://lists.ozlabs.org/pipermail/linuxppc-dev/2022-February/240254.html
Proposed Solution:
==================
Instead of reloading all kexec segments on hotplug event, this patch series
focuses on updating only the relevant kexec segment. Once the kexec segments
are loaded in the kernel reserved area then an arch-specific hotplug handler
will update the relevant kexec segment based on hotplug event type.
Series Dependecies
==================
This patch series implements the crash hotplug handler on PowerPC. The generic
crash hotplug update is introduced by https://lkml.org/lkml/2023/3/6/1358 patch
series.
Git tree for testing:
=====================
The below git tree has this patch series applied on top of dependent patch
series.
https://github.com/sourabhjains/linux/tree/in-kernel-crash-update-v9
To realise the feature the kdump udev rules must be disabled for CPU/Memory
hotplug events. Comment out the below line in kdump udev rule file:
RHEL: /usr/lib/udev/rules.d/98-kexec.rules
#SUBSYSTEM=="cpu", ACTION=="online", GOTO="kdump_reload_cpu"
#SUBSYSTEM=="memory", ACTION=="online", GOTO="kdump_reload_mem"
#SUBSYSTEM=="memory", ACTION=="offline", GOTO="kdump_reload_mem"
SLES: /usr/lib/kdump/70-kdump.rules
#SUBSYSTEM=="memory", ACTION=="add|remove", GOTO="kdump_try_restart"
#SUBSYSTEM=="cpu", ACTION=="online", GOTO="kdump_try_restart"
Sourabh,
The above seems to contradict what I anticipate to be udev rules changes once the base series is
accepted. Specifically I'm suggesting the following:
- Prevent udev from updating kdump crash kernel on hot un/plug changes.
Add the following as the first lines to the RHEL udev rule file
/usr/lib/udev/rules.d/98-kexec.rules:
# The kernel handles updates to crash elfcorehdr for cpu and memory changes
SUBSYSTEM=="cpu", ATTRS{crash_hotplug}=="1", GOTO="kdump_reload_end"
SUBSYSTEM=="memory", ATTRS{crash_hotplug}=="1", GOTO="kdump_reload_end"
With this changeset applied, the two rules evaluate to false for
cpu and memory change events and thus skip the userspace
unload-then-reload of kdump.
The above additions allow distros to deploy the udev rule immediately and work properly even if the
base patchset isn't yet merged, or down the road, enabled/configured.
Am I missing something such that your recommendation is different than mine?
Note: only kexec_file_load syscall will work. For kexec_load minor
changes are required in kexec tool.
Will this be the same/similar change as I have posted, or do you envision something different?
Thanks,
eric
---
Changelog:
v9:
- Removed patch to prepare elfcorehdr crash notes for possible CPUs.
The patch is moved to generic patch series that introduces generic
infrastructure for in kernel crash update.
- Removed patch to pass the hotplug action type to the arch crash
hotplug handler function. The generic patch series has introduced
the hotplug action type in kimage struct.
- Add detail commit message for better understanding.
v8:
- Restrict fdt_index initialization to machine_kexec_post_load
it work for both kexec_load and kexec_file_load.[3/8] Laurent Dufour
- Updated the logic to find the number of offline core. [6/8]
- Changed the logic to find the elfcore program header to accommodate
future memory ranges due memory hotplug events. [8/8]
v7
- added a new config to configure this feature
- pass hotplug action type to arch specific handler
v6
- Added crash memory hotplug support
v5:
- Replace COFNIG_CRASH_HOTPLUG with CONFIG_HOTPLUG_CPU.
- Move fdt segment identification for kexec_load case to load path
instead of crash hotplug handler
- Keep new attribute defined under kimage_arch to track FDT segment
under CONFIG_HOTPLUG_CPU config.
v4:
- Update the logic to find the additional space needed for hotadd CPUs post
kexec load. Refer "[RFC v4 PATCH 4/5] powerpc/crash hp: add crash hotplug
support for kexec_file_load" patch to know more about the change.
- Fix a couple of typo.
- Replace pr_err to pr_info_once to warn user about memory hotplug
support.
- In crash hotplug handle exit the for loop if FDT segment is found.
v3
- Move fdt_index and fdt_index_vaild variables to kimage_arch struct.
- Rebase patche on top of https://lkml.org/lkml/2022/3/3/674 [v5]
- Fixed warning reported by checpatch script
v2:
- Use generic hotplug handler introduced by https://lkml.org/lkml/2022/2/9/1406, a
significant change from v1.
Sourabh Jain (6):
powerpc/kexec: turn some static helper functions public
powerpc/crash: introduce a new config option CRASH_HOTPLUG
powerpc/crash: add a new member to the kimage_arch struct
powerpc/crash: add crash CPU hotplug support
crash: forward memory_notify args to arch crash hotplug handler
powerpc/kexec: add crash memory hotplug support
arch/powerpc/Kconfig | 12 +
arch/powerpc/include/asm/kexec.h | 15 ++
arch/powerpc/include/asm/kexec_ranges.h | 1 +
arch/powerpc/kexec/core_64.c | 322 ++++++++++++++++++++++++
arch/powerpc/kexec/elf_64.c | 13 +-
arch/powerpc/kexec/file_load_64.c | 212 ++++------------
arch/powerpc/kexec/ranges.c | 85 +++++++
arch/x86/include/asm/kexec.h | 2 +-
arch/x86/kernel/crash.c | 3 +-
include/linux/kexec.h | 2 +-
kernel/crash_core.c | 14 +-
11 files changed, 506 insertions(+), 175 deletions(-)
_______________________________________________
kexec mailing list
kexec@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/kexec
