All of lore.kernel.org
 help / color / mirror / Atom feed
From: mwilck@suse.com
To: Christophe Varoqui <christophe.varoqui@opensvc.com>,
	Benjamin Marzinski <bmarzins@redhat.com>
Cc: dm-devel@redhat.com, Martin Wilck <mwilck@suse.com>
Subject: [dm-devel] [PATCH 0/7] Various multipath-tools patches
Date: Thu, 17 Dec 2020 12:00:11 +0100	[thread overview]
Message-ID: <20201217110018.3347-1-mwilck@suse.com> (raw)

From: Martin Wilck <mwilck@suse.com>

Hi Christophe, hi Ben,

here are a few patches that have accumulated while I hadn't
fixed my previous series "libmultipath: improve cleanup on exit"

The first two are actually related to the log_thread work I did
in that series, but they're additional fixes on top and I didn't
want to delay that series further by adding additional, possibly
controversial stuff.

No 3 is the 2nd attempt to allow unloading of checker DSOs, hopefully
fixing the issues Ben saw with the first one.

No 4 is probably the most important one. It fixes the same issue
that 2b25a9e and cb10d38 targeted, but the previous fixes still
didn't cover all corner cases. This one does, I believe.

No 5 tackles a new failure mode we recently discovered - multipathd
may block trying to access the root FS while this is in queuing mode.
This patch removes one very frequent access pattern.
This is work in progress; there are other file system accesses
from multipathd which might block.

Obviously the wwids, prkeys, and bindings files' contents 
need to be cached. I haven't made up my mind how to handle writing
changes back. Would be sufficient to write back when multipathd exits,
possibly with an external program (e.g. "ExecStartPost")? Or should
we try to write back directly at runtime, using separate threads to
avoid blocking of the main thread? "reconfigure" is another issue,
but perhaps less problematic - if the root FS is blocked, users
couldn't edit /etc/multipath.conf anyway. One might consider moving
everything to tmpfs or even running in a tmpfs-backed chroot, but
it still would beg the question how to make certain changes permanent,
like wwids and bindings.

6 and 7 fix oversights in my library versioning series. The
plugins didn't have symbol versions set on the symbols they referenced,
making the versioning at least partly non-functional.

Regards
Martin

Martin Wilck (7):
  libmultipath: move logq_lock handling to log.c
  libmultipath: protect logarea with logq_lock
  libmultipath: prevent DSO unloading with astray checker threads
  libmultipath: force map reload if udev incomplete
  multipath-tools: avoid access to /etc/localtime
  multipath-tools: make sure plugin DSOs use symbol versions
  libmultipath.version: add missing symbol

 Makefile                           |  1 +
 Makefile.inc                       |  2 +-
 libmpathpersist/Makefile           |  8 ++--
 libmultipath/checkers.c            | 68 +++++++++++++++++++++++++++---
 libmultipath/checkers.h            | 25 +++++++++++
 libmultipath/checkers/Makefile     |  7 ++-
 libmultipath/checkers/tur.c        | 12 +++---
 libmultipath/configure.c           | 45 +++++++++++++-------
 libmultipath/debug.c               | 14 +++---
 libmultipath/devmapper.c           | 12 +++---
 libmultipath/foreign/Makefile      |  4 +-
 libmultipath/libmultipath.version  | 10 +++++
 libmultipath/log.c                 | 59 ++++++++++++++++++++++----
 libmultipath/log.h                 |  1 -
 libmultipath/log_pthread.c         |  9 ----
 libmultipath/prioritizers/Makefile |  7 ++-
 multipathd/main.c                  |  3 --
 17 files changed, 211 insertions(+), 76 deletions(-)

-- 
2.29.0


--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel


             reply	other threads:[~2020-12-17 11:01 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-12-17 11:00 mwilck [this message]
2020-12-17 11:00 ` [dm-devel] [PATCH 1/7] libmultipath: move logq_lock handling to log.c mwilck
2020-12-17 23:56   ` Benjamin Marzinski
2020-12-17 11:00 ` [dm-devel] [PATCH 2/7] libmultipath: protect logarea with logq_lock mwilck
2020-12-18  0:03   ` Benjamin Marzinski
2020-12-18 16:24     ` Martin Wilck
2020-12-18 16:32       ` Benjamin Marzinski
2020-12-17 11:00 ` [dm-devel] [PATCH 3/7] libmultipath: prevent DSO unloading with astray checker threads mwilck
2020-12-18  4:22   ` Benjamin Marzinski
2020-12-17 11:00 ` [dm-devel] [PATCH 4/7] libmultipath: force map reload if udev incomplete mwilck
2020-12-18  5:48   ` Benjamin Marzinski
2020-12-18 15:06     ` Martin Wilck
2020-12-18 15:08       ` Martin Wilck
2020-12-18 17:57   ` Benjamin Marzinski
2020-12-17 11:00 ` [dm-devel] [PATCH 5/7] multipath-tools: avoid access to /etc/localtime mwilck
2020-12-18 18:14   ` Benjamin Marzinski
2020-12-17 11:00 ` [dm-devel] [PATCH 6/7] multipath-tools: make sure plugin DSOs use symbol versions mwilck
2020-12-18 18:36   ` Benjamin Marzinski
2020-12-17 11:00 ` [dm-devel] [PATCH 7/7] libmultipath.version: add missing symbol mwilck
2020-12-18 18:36   ` Benjamin Marzinski

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20201217110018.3347-1-mwilck@suse.com \
    --to=mwilck@suse.com \
    --cc=bmarzins@redhat.com \
    --cc=christophe.varoqui@opensvc.com \
    --cc=dm-devel@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.