Hi, Kairui
Thank you for the improvement.
This series looks good to me. But I noticed some typos(I put them at the end) in the patch
log,
you could correct them when merging this series, and no need to post again.
Acked-by: Lianbo Jiang <lijiang(a)redhat.com>
Lianbo
在 2020年03月12日 21:33, Kairui Song 写道:
With this series, bind mount will be treated equally, and no longer
have
special workaround for atomic. And also cleaned up user configured
target handling code. There should be no function change by this fat.
Another fix applied here is that in atomic/silverblue enviroment, the
path is always resolved as current bind mounted path under root device,
but for user configured targets, path should be just used as it is else
the behavior will be uncontrollable and go wrong easily.
Kairui Song (5):
Remove is_dump_target_configured
mkdumprd: Use get_save_path instead of parsing config
mkdumprd: Simplify handling of user specified target
No longer treat atomic/silverblue specially
Remove adjust_bind_mount_path call
dracut-module-setup.sh | 39 +---------------
kdump-lib.sh | 102 +++++++++++++----------------------------
kdumpctl | 18 ++------
mkdumprd | 37 ++++-----------
4 files changed, 48 insertions(+), 148 deletions(-)
Some typos in patch log:
[1] Current code logic is not only complicated, it also wrongly generate
an redundantly long path in atomic/silverblue ***enviroment***.
[2] The target is large ***engouth*** to hold to contain the vmcore.
[3] This commit remove (***almose***) all special workaround for atomic, and treat
[4] all bind mounts in any ***enviroment*** equally.
[5] Use a ***heleper*** get_bind_mount_directory_from_path to get the bind mount
[6] for atomic/silverblue ***enviroment***.