From: Daniel Wagner <wagi@kernel.org>
To: James Smart <james.smart@broadcom.com>,
Christoph Hellwig <hch@lst.de>, Sagi Grimberg <sagi@grimberg.me>,
Chaitanya Kulkarni <kch@nvidia.com>
Cc: Hannes Reinecke <hare@suse.de>, Keith Busch <kbusch@kernel.org>,
linux-nvme@lists.infradead.org, linux-kernel@vger.kernel.org,
Daniel Wagner <wagi@kernel.org>
Subject: [PATCH v3 00/18] nvmet-fcloop: track resources via reference counting
Date: Tue, 18 Mar 2025 11:39:54 +0100 [thread overview]
Message-ID: <20250318-nvmet-fcloop-v3-0-05fec0fc02f6@kernel.org> (raw)
It turns out v2 had a memory leak, which sent me down the rabbit hole
once again. The obvious leak was easy to fix. I took one reference too
much on fcloop_lsreq.
The second one was a really tricky. The fcloop_lsreq object are allocated
by either host or target. So the lifetime of these object are tied to the
lifetime of the host or target. I tried several things to avoid UAFs but
everything failed. Eventually, I decided to allocated the fcloop_lsreqs
in fcloop directly. All got way simpler and the mem leak is gone and
there are no UAFs anymore or blocked processes.
Sometimes some blktests fail but these really look like existing problems
which got uncovered by this series.
With this series the fcloop and nvmet-fc should be in way better shape
and hopefully most of the UAFs should be history. This allows futher
refactoring/cleanup/improvements or more nasty blktests.
Signed-off-by: Daniel Wagner <wagi@kernel.org>
---
Changes in v3:
- fixed memory leaks
- allocates fcloop_lsreq in fcloop directly
- prevent double free due to unregister race
- collected tags
- Link to v2: https://lore.kernel.org/r/20250311-nvmet-fcloop-v2-0-fc40cb64edea@kernel.org
Changes in v2:
- drop tport and rport ref counting, use implicit synchronisation
- a bunch of additional fixes in existing ref countig
- replaced kref with refcount
- Link to v1: https://lore.kernel.org/r/20250226-nvmet-fcloop-v1-0-c0bd83d43e6a@kernel.org
---
Daniel Wagner (18):
nvmet-fcloop: remove nport from list on last user
nvmet-fcloop: replace kref with refcount
nvmet-fcloop: add ref counting to lport
nvmet-fcloop: refactor fcloop_nport_alloc
nvmet-fcloop: track ref counts for nports
nvmet-fcloop: sync targetport removal
nvmet-fcloop: update refs on tfcp_req
nvmet-fcloop: add missing fcloop_callback_host_done
nvmet-fcloop: prevent double port deletion
nvmet-fcloop: allocate/free fcloop_lsreq directly
nvmet-fc: inline nvmet_fc_delete_assoc
nvmet-fc: inline nvmet_fc_free_hostport
nvmet-fc: update tgtport ref per assoc
nvmet-fc: take tgtport reference only once
nvmet-fc: free pending reqs on tgtport unregister
nvmet-fc: take tgtport refs for portentry
nvmet-fc: put ref when assoc->del_work is already scheduled
nvme-fc: do not reference lsrsp after failure
drivers/nvme/host/fc.c | 13 +-
drivers/nvme/target/fc.c | 153 +++++++++------
drivers/nvme/target/fcloop.c | 435 ++++++++++++++++++++++++++-----------------
3 files changed, 376 insertions(+), 225 deletions(-)
---
base-commit: a149420f548dc8e2258461522f498252554c0bbd
change-id: 20250214-nvmet-fcloop-a649738b7e6e
Best regards,
--
Daniel Wagner <wagi@kernel.org>
next reply other threads:[~2025-03-18 10:40 UTC|newest]
Thread overview: 62+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-03-18 10:39 Daniel Wagner [this message]
2025-03-18 10:39 ` [PATCH v3 01/18] nvmet-fcloop: remove nport from list on last user Daniel Wagner
2025-03-18 10:39 ` [PATCH v3 02/18] nvmet-fcloop: replace kref with refcount Daniel Wagner
2025-03-18 10:56 ` Hannes Reinecke
2025-04-02 14:03 ` Daniel Wagner
2025-04-02 14:06 ` Hannes Reinecke
2025-03-21 6:03 ` Christoph Hellwig
2025-03-21 14:16 ` Hannes Reinecke
2025-03-18 10:39 ` [PATCH v3 03/18] nvmet-fcloop: add ref counting to lport Daniel Wagner
2025-03-21 6:04 ` Christoph Hellwig
2025-03-18 10:39 ` [PATCH v3 04/18] nvmet-fcloop: refactor fcloop_nport_alloc Daniel Wagner
2025-03-18 11:02 ` Hannes Reinecke
2025-03-18 13:38 ` Daniel Wagner
2025-03-18 14:10 ` Hannes Reinecke
2025-03-21 6:05 ` Christoph Hellwig
2025-03-18 10:39 ` [PATCH v3 05/18] nvmet-fcloop: track ref counts for nports Daniel Wagner
2025-03-18 11:06 ` Hannes Reinecke
2025-03-18 13:44 ` Daniel Wagner
2025-03-18 10:40 ` [PATCH v3 06/18] nvmet-fcloop: sync targetport removal Daniel Wagner
2025-03-18 11:07 ` Hannes Reinecke
2025-03-21 6:08 ` Christoph Hellwig
2025-04-02 16:39 ` Daniel Wagner
2025-03-18 10:40 ` [PATCH v3 07/18] nvmet-fcloop: update refs on tfcp_req Daniel Wagner
2025-03-18 11:09 ` Hannes Reinecke
2025-03-21 6:08 ` Christoph Hellwig
2025-03-18 10:40 ` [PATCH v3 08/18] nvmet-fcloop: add missing fcloop_callback_host_done Daniel Wagner
2025-03-18 11:12 ` Hannes Reinecke
2025-03-18 13:49 ` Daniel Wagner
2025-04-02 17:08 ` Daniel Wagner
2025-04-03 13:25 ` Daniel Wagner
2025-04-04 7:28 ` Daniel Wagner
2025-03-21 6:12 ` Christoph Hellwig
2025-03-18 10:40 ` [PATCH v3 09/18] nvmet-fcloop: prevent double port deletion Daniel Wagner
2025-03-18 11:15 ` Hannes Reinecke
2025-03-18 13:55 ` Daniel Wagner
2025-03-21 6:13 ` Christoph Hellwig
2025-03-18 10:40 ` [PATCH v3 10/18] nvmet-fcloop: allocate/free fcloop_lsreq directly Daniel Wagner
2025-03-18 11:17 ` Hannes Reinecke
2025-03-18 13:58 ` Daniel Wagner
2025-04-08 11:20 ` Daniel Wagner
2025-03-19 22:47 ` James Smart
2025-04-04 12:53 ` Daniel Wagner
2025-03-18 10:40 ` [PATCH v3 11/18] nvmet-fc: inline nvmet_fc_delete_assoc Daniel Wagner
2025-03-21 6:14 ` Christoph Hellwig
2025-03-18 10:40 ` [PATCH v3 12/18] nvmet-fc: inline nvmet_fc_free_hostport Daniel Wagner
2025-03-21 6:15 ` Christoph Hellwig
2025-03-18 10:40 ` [PATCH v3 13/18] nvmet-fc: update tgtport ref per assoc Daniel Wagner
2025-03-21 6:15 ` Christoph Hellwig
2025-03-18 10:40 ` [PATCH v3 14/18] nvmet-fc: take tgtport reference only once Daniel Wagner
2025-03-18 11:18 ` Hannes Reinecke
2025-03-21 6:17 ` Christoph Hellwig
2025-03-18 10:40 ` [PATCH v3 15/18] nvmet-fc: free pending reqs on tgtport unregister Daniel Wagner
2025-03-21 6:19 ` Christoph Hellwig
2025-04-08 11:29 ` Daniel Wagner
2025-03-18 10:40 ` [PATCH v3 16/18] nvmet-fc: take tgtport refs for portentry Daniel Wagner
2025-03-18 11:19 ` Hannes Reinecke
2025-03-18 10:40 ` [PATCH v3 17/18] nvmet-fc: put ref when assoc->del_work is already scheduled Daniel Wagner
2025-03-18 11:20 ` Hannes Reinecke
2025-03-21 6:19 ` Christoph Hellwig
2025-03-18 10:40 ` [PATCH v3 18/18] nvme-fc: do not reference lsrsp after failure Daniel Wagner
2025-03-18 11:33 ` Hannes Reinecke
2025-03-18 14:01 ` Daniel Wagner
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=20250318-nvmet-fcloop-v3-0-05fec0fc02f6@kernel.org \
--to=wagi@kernel.org \
--cc=hare@suse.de \
--cc=hch@lst.de \
--cc=james.smart@broadcom.com \
--cc=kbusch@kernel.org \
--cc=kch@nvidia.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-nvme@lists.infradead.org \
--cc=sagi@grimberg.me \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®