Commit Graph

1211 Commits

Author SHA1 Message Date
Kairui Song
978a849765 selftest: Add document for selftests
Acked-by: Dave Young <dyoung@redhat.com>
2020-09-17 10:42:58 +08:00
Kairui Song
a8dbd281f7 selftest: Add basic test framework
Now, by execute `make test-run` in tests/, kexec-tools sanity tests
will be performed using VMs. There are currently 3 test cases, for local
kdump, nfs kdump and ssh kdump.

For each test VM, the selftest framework will create a snapshot layer,
do setup as required by the test case, this ensure each test runs in a
clean VM.

This framework will install a custom systemd service that starts when
system have finished booting, and the service will do basic routine
(fetch and set boot counter, etc..), then call the test case which is
installed in /kexec-kdump-test/test.sh in VM.

Each VM will have two serial consoles, one for ordinary console usage,
one for the test communication and log. The test script will watch the
second test console to know the test status.

The test cases are located in tests/scripts/testcases, documents about
the test cases structure will be provided in following commits.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-09-17 10:42:54 +08:00
Kairui Song
2457f22baf selftest: Add basic infrastructure to build test image
The Makefile In tests/ could help build a VM image using Fedora cloud
image as base image, or, user can specify a base image using
BASE_IMAGE=<path/to/file>. The current repo will be packeged and
installed in the image, so the image could be used as a test image to
test kexec-tools.

The image building is splited into two steps:
The first step, it either convert the base image to qcow2 or create
a snapshot on it, and install basic packages (dracut, grubby, ...)
and do basic setups (setup crashkernel=, disable selinux, ...).
See tests/scripts/build-scripts/base-image.sh for detail.

The second step, it creates a snapshot on top of the image produced by
the previous step, and install the packaged kexec-tools of current
repo.  See tests/scripts/build-scripts/test-base-image.sh for detail.

In this way, if repo's content is changes, `make` will detect it and
only rebuild the second snapshot which speed up the rebuild by a lot.

The image will be located as tests/output/test-base-image, and in qcow2
format. And default user/password is set to root/fedora.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-09-17 10:42:34 +08:00
Kairui Song
3d5d9f0c0f Release 2.0.20-18
Signed-off-by: Kairui Song <kasong@redhat.com>
2020-08-27 13:14:59 +08:00
Kairui Song
5ec98c2ec7 mkdumprd: Improve the warning message when using encrypted target
Encrypted target have many issues, so let user check
kexec-kdump-howto.txt, which have more details.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-08-27 13:10:42 +08:00
Kairui Song
ff46cfb19e kdump-lib.sh: Remove is_atomic
Now all atomic special workaround is removed, we can remove the atomic
detection function.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-08-27 11:35:01 +08:00
Kairui Song
a37f36ad4d Refactor kernel image and initrd detection code
kernel installation is not always in a fixed location /boot, there are
multiple different style of kernel installation, and initramfs location
changes with kernel. The two files should be detected together and adapt
to different style.

To do so we use a list of known installation destinations, and a list
of possible kernel image and initrd names. Iterate the two list to
detect the installation location of the two files. If GRUB is in use,
the BOOT_IMAGE= cmdline from GRUB will also be considered. And also
prefers user specified config if given.

Previous atomic workaround is no longer needed as the new detection
method can cover that case.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-08-27 11:29:17 +08:00
Kairui Song
c76820bddd early-kdump: Use consistent symbol link for kernel and initramfs
There is no need to follow user's configuration when installtion the
kernel and initramfs for early kdump, just use a fixed a symbol link is
enough, this help avoid a check_boot_dir call.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-08-27 11:23:45 +08:00
Kairui Song
a7199c141c Add a kernel install hook to clean up kdump initramfs
Kdump service will create kdump initramfs when needed, but it won't
clean up the kdump initramfs on kernel uninstall. So create a kernel
install hook to do the clean up job.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-08-27 11:22:03 +08:00
Kairui Song
f8f8d79d7f Release 2.0.20-17
Signed-off-by: Kairui Song <kasong@redhat.com>
2020-08-04 10:50:46 +08:00
Kairui Song
d0d9ff1eb2 Drop static lib dependencies
Signed-off-by: Kairui Song <kasong@redhat.com>
2020-08-04 10:47:51 +08:00
Kairui Song
073c30973c Revert "x86_64: enable the kexec file load by default"
This reverts commit e5cbf4d1ad.

Currently the kernel image is being dual signed, but kexec_file_load
can't handle multiple signature right, so it will always fail. Let's
temporaryly revert "using '-s' by default", until upstream have fixed
the real issue.

Signed-off-by: Kairui Song <kasong@redhat.com>
2020-08-04 10:47:51 +08:00
Kairui Song
fa8aa52d94 Revert "s390x: enable the kexec file load by default"
This reverts commit 8884af04b6.

Currently the kernel image is being dual signed, but kexec_file_load
can't handle multiple signature right, so it will always fail. Let's
temporaryly revert "using '-s' by default", until upstream have fixed
the real issue.

Signed-off-by: Kairui Song <kasong@redhat.com>
2020-08-04 10:47:51 +08:00
Fedora Release Engineering
21b4996e3a - Second attempt - Rebuilt for https://fedoraproject.org/wiki/Fedora_33_Mass_Rebuild
Signed-off-by: Fedora Release Engineering <releng@fedoraproject.org>
2020-08-01 03:43:49 +00:00
Pingfan Liu
f96172d353 kdumpctl: exit if either pre.d or post.d is missing
It is hard to detect the time that /etc/kdump is removed. And this failure
may cause out-of-date kdump.initrd.  To keep things simple, just exit if
/etc/kdump/pre.d and post.d does not exist.

Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Kairui Song <kasong@redhat.com>
2020-07-30 16:47:10 +08:00
Fedora Release Engineering
1844d088f7 - Rebuilt for https://fedoraproject.org/wiki/Fedora_33_Mass_Rebuild
Signed-off-by: Fedora Release Engineering <releng@fedoraproject.org>
2020-07-28 03:00:49 +00:00
Pingfan Liu
bc67c13651 kdump_pre: make notes more precise
Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Kairui Song <kasong@redhat.com>
2020-07-21 10:34:36 +08:00
Pingfan Liu
bda677c3d1 dracut-kdump.sh: exit shell when machine reboot
The following scenario is observed:

  kdump: kdump_pre script exited with non-zero status!
  [    5.104841] systemd[1]: Shutting down.
  [    5.122162] printk: systemd-shutdow: 27 output lines suppressed due to ratelimiting
  kdump: dump target is /dev/mapper/rhel_hpe--dl380pgen8--02--vm--12-root
  kdump: saving to /sysroot//var/crash/127.0.0.1-2020-06-27-03:55:01/
  kdump: saving vmcore-dmesg.txt
  kdump: saving vmcore-dmesg.txt complete
  kdump: saving vmcore
   Checking for memory holes                         : [  0.0 %] /                   Checking for memory holes                         : [100.0 %] |                  [    5.516573] systemd-shutdown[1]: Syncing filesystems and block devices.
  [    5.519515] systemd-shutdown[1]: Sending SIGTERM to remaining processes...

It is caused by the following script
  if [ $? -ne 0 ]; then
      echo "kdump: kdump_pre script exited with non-zero status!"
      do_final_action
  fi

When do_final_action runs, a systemd service is forked for reboot, then the
subshell returns, and parent continues to execute. Place "exit 1" to stop
executing and make kdump service failure.

Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Kairui Song <kasong@redhat.com>
2020-07-21 10:34:07 +08:00
Pingfan Liu
25824d64cd kdumpctl: detect modification of scripts by its directory's timestamp
Checking modification against a file can not detect a removing file in
"/etc/kdump/post.d/ /etc/kdump/pre.d/".  Hence it also needs the
modified time of directory to detect such changes.

Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Kairui Song <kasong@redhat.com>
2020-07-20 16:18:42 +08:00
Pingfan Liu
bdbddbff73 module-setup.sh: suppress false alarm
Even if the directory "/etc/kdump/pre.d/" is empty, the following false
alarm can be observed during building kdump.initrd:
  "/etc/kdump/pre.d/* is not executable"

Suppress it.

Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Kairui Song <kasong@redhat.com>
2020-07-20 16:17:53 +08:00
Pingfan Liu
8bef2d943b kexec-tools.spec: make the existence of pre.d and post.d mandatory
If the directory /etc/kdump/pre.d is optional, then it is hard
to tell between the following cases:
  -1. no directory "/etc/kdump/pre.d"
  -2. "rm -rf /etc/kdump/pre.d", which removes all scripts under pre.d
For the second case, kdump.img should be rebuilt.

To bail out from this corner case, always creating pre.d and post.d
during rpm installation.

Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Kairui Song <kasong@redhat.com>
2020-07-20 16:12:04 +08:00
Hari Bathini
5521a70cc7 ppc64/kdump: use kexec_file_load when secureboot is enabled
In secure boot mode, kexec_load syscall is disabled. So, if in secure
boot mode, load kdump kernel with kexec_file_load syscall instead.

Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-07-20 15:21:54 +08:00
Kairui Song
070d362846 Release 2.0.20-14
Also fix a timestamp issue.

Signed-off-by: Kairui Song <kasong@redhat.com>
2020-07-01 17:15:38 +08:00
Lianbo Jiang
8884af04b6 s390x: enable the kexec file load by default
Let's use the option KEXEC_ARGS="-s" to enable the kexec file load
by default.

Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-07-01 17:07:46 +08:00
Lianbo Jiang
e5cbf4d1ad x86_64: enable the kexec file load by default
Let's use the option KEXEC_ARGS="-s" to enable the kexec file load
by default.

Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-07-01 17:07:46 +08:00
Lianbo Jiang
bf20a948ef Revert "s390x: add kdump sysconfig option to use the kexec_file_load() syscall"
This reverts commit 66ff48ed68.

Since the code logic is changed(which related to the option
'KDUMP_FILE_LOAD=on|off'), accordingly, need to remove the
option from the sysconfig file.

Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-07-01 17:07:46 +08:00
Lianbo Jiang
073646998f Revert "kdump-lib: switch to the kexec_file_load() syscall on x86_64 by default"
This reverts commit 6a20bd5447.

Let's restore the logic of secureboot status check, and remove the
option 'KDUMP_FILE_LOAD=on|off'. We will use the option KEXEC_ARGS="-s"
to enable the kexec file load later, which can avoid failures when
the secureboot is enabled.

Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-07-01 17:07:46 +08:00
Pingfan Liu
f2824e7b74 kdump.conf: fix a grammar issue
Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-07-01 17:07:03 +08:00
Pingfan Liu
6dd2b9fb41 man: improve description about /etc/kdump/{pre.d,post.d}interface
The description is identical to the notes in kdump.conf

Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Kairui Song <kasong@redhat.com>
2020-07-01 16:49:41 +08:00
Kairui Song
61d4fa72b5 mkdumprd: Improve the error message for umounted dump target
Since commit ca01cbd, mkdumprd can generate the initramfs even if dump
target is not mounted. But it will refuse to work if the dump target is
not marked as noauto, and not mounted.

If a target is not marked as noauto in fstab, it should be mounted
automatically on boot. If it's not mounted, most likely something is
wrong and user should fix that. So mkdumprd will just show an error
message and quit to avoid further failure.

This commit make the error message more clear and less confusing.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-06-24 14:15:49 +08:00
Kairui Song
807f0b5e48 mkdumprd: Fix nfs detection in to_mount
If target is in format of UUID= or LABEL=, current code will fail as it
expects target to be the absoluta path to the block device. Check for
the fstype instead.

Also simplify the code logic.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-06-23 09:34:53 +08:00
Kairui Song
a29de38da5 Always wrap up call to dracut get_persistent_dev function
Dracut get_persistent_dev function don't recognize UUID= or LABEL=
format, so caller should conver it to the path to the block device
before calling it. There is already such a helper
"kdump_get_persistent_dev", just move it to kdump-lib.sh and rename
it to reuse it,

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-06-22 19:58:08 +08:00
Lianbo Jiang
66ff48ed68 s390x: add kdump sysconfig option to use the kexec_file_load() syscall
Currently, kernel has supported the kexec_file_load() syscall on s390x,
so I follow this commit:
6a20bd5447 ("kdump-lib: switch to the kexec_file_load() syscall on
x86_64 by default")

And also enable the kexec_file_load() syscall for kdump on s390x by
default.

Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
Acked-by: Kairui Song <kasong@redhat.com>
2020-06-15 18:20:20 +08:00
Kairui Song
5d4a5896c9 mkdumprd: Fix dracut error on multiple extra_modules
Dracut expects --add-drivers with a single parameter, so mkdumprd
should quota the module list.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-06-15 17:42:35 +08:00
Kairui Song
1a5d44d7f4 Fix kdump failure when mount target specified by dracut_args
commit 61e0169 changed definition of dump_fs function, so
need to do a mount target conversion before calling it.

Signed-off-by: Kairui Song <kasong@redhat.com>
2020-06-11 14:00:54 +08:00
onitsuka.shinic@fujitsu.com
077a7f19bb kdump.conf: Specify /etc/kdump/{pre.d,post.d}interface
This patch discribe /etc/kdump/{pre.d,post.d} interface.

Signed-off-by: Shinichi Onitsuka <onitsuka.shinic@fujitsu.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-06-11 13:18:26 +08:00
onitsuka.shinic@fujitsu.com
45e02e73fa dracut-kdump.sh: Execute the binary and script filesin /etc/kdump/{pre.d,post.d}
This patch executes the binary and script files in /etc/kdump/{pre.d,post.d}
just like kdump_pre or kdump_post directive written in /etc/kdump.conf.

Signed-off-by: Shinichi Onitsuka <onitsuka.shinic@fujitsu.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-06-11 12:59:21 +08:00
onitsuka.shinic@fujitsu.com
bdd57a5864 kdumpctl: Check the update of the binary and script files in /etc/kdump/{pre.d,post.d}
This patch adds the binary and script files in /etc/kdump/{pre.d,post.d}
to modified checklist in order to update kdump initramfs when one adds
new scripts or binaries or removes the existing ones under
/etc/kdump/{pre.d, post.d}.

Signed-off-by: Shinichi Onitsuka <onitsuka.shinic@fujitsu.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-06-11 12:59:15 +08:00
onitsuka.shinic@fujitsu.com
4246f26725 dracut-module-setup.sh: Install files under /etc/kdump/{pre.d,post.d} into kdump initramfs
This patch installs the binary and script files under /etc/kdump/{pre.d,post.d}
into new initramfs of kdump.

Signed-off-by: Shinichi Onitsuka <onitsuka.shinic@fujitsu.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-06-11 12:58:48 +08:00
Kairui Song
e05c550144 Drop switch root capability for non fadump initramfs
Switch root is never used for kdump image, and this will be helpful to
reduce the initramfs size.

Also increase dracut dependency version and the function is
dracut_no_switch_root is new introduced.

This commit is applied to RHEL some time ago, but missing in Fedora as
Fedora's Dracut didn't backport this feature at that time. Now apply
this missing commit.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-06-10 22:37:34 +08:00
Hari Bathini
8fcbb4d265 fadump: update fadump-howto.txt with some more troubleshooting help
On journaling filesystems like XFS, bootloader is likely to pick up
older initrd (without vmcore capture scripts) if system crashes right
after initrd update, as the bootloader (read GRUB) may not replay
filesystem log before reading the initrd from disk. Added steps to
workaround that problem.

Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-06-04 11:45:05 +08:00
Pingfan Liu
47bdeee35d fadump-howto.txt: source it in spec file
Signed-off-by: Pingfan Liu <piliu@redhat.com>
Acked-by: Kairui Song <kasong@redhat.com>
2020-06-03 11:34:37 +08:00
Kairui Song
134c578463 Don't inherit swiotlb parameter form 1st kernel by default
Usually kdump kernel don't need a swiotlb. There are two cases known
common case a swiotlb is needed:

 - SEV/SME is active.
 - crashkernel high reservation is used, or there are crashkernel memory
   above 4G.

For both case, if swiotlb is set to a large value in first kernel, kdump
kernel will likely to fail with OOM. So don't inherit swiotlb parameter
by default.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-05-29 10:56:31 +08:00
Kairui Song
0cc3b85d0d module-setup.sh: Add "rd.neednet" parameter if network is needed
Upstream dracut now use network-manager module by default and since
upstream commit 3dcaa97, network-manager expects user to pass "rd.neednet"
to indicate network is required.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-05-28 16:26:06 +08:00
Kairui Song
cfd93e2b7e Revert "Add a hook to wait for kdump target in initqueue"
This reverts commit cee618593c.

Upstream dracut have provided a parameter for adding mandantory network
requirement by appending "rd.neednet" parameter, so we should use that
instead.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
2020-05-28 16:26:00 +08:00
Lianbo Jiang
a81ec90ed8 kdump.sysconfig: Remove the option 'log_buf_len' from kdump command line
The option 'log_buf_len' in the kernel command line will make kernel
dynamically allocate memory (system ram) from memblock and always hold
the memory allocated. If the size of this option is misused such as
'log_buf_len=64M'. This may cause a short of memory for kdump kernel,
because the size of memory is usually limited in kdump kernel.

In order to avoid this problem, need to remove this option from the
kdump kernel command line.

Signed-off-by: Lianbo Jiang <lijiang@redhat.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-05-28 10:11:15 +08:00
Kairui Song
0094fb496d Release 2.0.20-13
Add support for using unmounted device as dump target, update docs, and
fix several known issue.

Signed-off-by: Kairui Song <kasong@redhat.com>
2020-05-22 19:29:19 +08:00
Kairui Song
73da3224be Fix an error in commit ca01cbd
Previous commit ca01cbd ('mkdumprd: generate usable kdump initramfs
even target is not mounted') applied a wrong version of patch, let me
just fix it.

Signed-off-by: Kairui Song <kasong@redhat.com>
2020-05-22 19:29:19 +08:00
Kairui Song
846e23a2d5 Update docs for the new noauto dump target support
Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-05-22 16:16:01 +08:00
Kairui Song
f2ffa3d340 kexec-kdump-howto.txt: Add some format to the document
When adding doc for the non-mounted dump target support, I found the
document are a bit uneasy to read due to lack of a proper format, this
commit should it make looks better.

Signed-off-by: Kairui Song <kasong@redhat.com>
Acked-by: Pingfan Liu <piliu@redhat.com>
2020-05-22 16:14:13 +08:00