From: "Aurélien Aptel" <aaptel@suse.com>
To: Paul Aurich <paul@darkrain42.org>
Cc: linux-cifs@vger.kernel.org, linux-sparse@vger.kernel.org,
sfrench@samba.org, Ronnie Sahlberg <lsahlber@redhat.com>
Subject: lock checking issues (was: Re: [PATCH v3] cifs: Fix leak when handling lease break for cached root fid)
Date: Tue, 07 Jul 2020 09:03:17 +0200 [thread overview]
Message-ID: <87tuyj6c3u.fsf@suse.com> (raw)
In-Reply-To: <20200706192642.GA110607@haley.home.arpa>
CC-ing linux-sparse
We are seeing a lock context imbalance warning which we can't get rid
of after applying a patch moving locking around across function boundaries.
For context see:
https://lore.kernel.org/linux-cifs/20200702164411.108672-1-paul@darkrain42.org/T/#u
Paul Aurich <paul@darkrain42.org> writes:
> On 2020-07-06 10:30:27 +0200, Aurélien Aptel wrote:
>>Paul Aurich <paul@darkrain42.org> writes:
>>> Changes since v2:
>>> - address sparse lock warnings by inlining smb2_tcon_has_lease and
>>> hardcoding some return values (seems to help sparse's analysis)
>>
>>Ah, I think the issue is not the inlining but rather you need to
>>instruct sparse that smb2_tcon_hash_lease is expected to release the
>>lock.
>>
>>https://www.kernel.org/doc/html/v4.12/dev-tools/sparse.html#using-sparse-for-lock-checking
>>
>>Probably with __releases somewhere in the func prototype.
>
> I tried various iterations of that without finding one that seems to work.
> I suspect it's because the unlocking is _conditional_.
Hm could be it...
> w/o the inline and with __releases(&cifs_tcp_ses_lock):
>
> fs/cifs/smb2misc.c:508:1: warning: context imbalance in 'smb2_tcon_has_lease'
> - different lock contexts for basic block
> fs/cifs/smb2misc.c:612:25: warning: context imbalance in
> 'smb2_is_valid_lease_break'- different lock contexts for basic block
>
>
> Digging further, I found __acquire and __release (not plural), which can be
> used in individual blocks. The following seems to silence the sparse warnings
> - does this seem valid?
To be honnest I'm not sure, these seem counterproductive. If you are
indicating you are acquiring X but lock Y the next line it feels like we
are fighting the tool instead of letting it help us.
> @@ -504,7 +504,7 @@ cifs_ses_oplock_break(struct work_struct *work)
> kfree(lw);
> }
>
> -static inline bool
> +static bool
> smb2_tcon_has_lease(struct cifs_tcon *tcon, struct smb2_lease_break *rsp)
> {
> bool found;
> @@ -521,6 +521,7 @@ smb2_tcon_has_lease(struct cifs_tcon *tcon, struct smb2_lease_break *rsp)
>
> lease_state = le32_to_cpu(rsp->NewLeaseState);
>
> + __acquire(cifs_tcp_ses_lock);
Should it have a "&" here?
Cheers,
--
Aurélien Aptel / SUSE Labs Samba Team
GPG: 1839 CB5F 9F5B FB9B AA97 8C99 03C8 A49B 521B D5D3
SUSE Software Solutions Germany GmbH, Maxfeldstr. 5, 90409 Nürnberg, DE
GF: Felix Imendörffer, Mary Higgins, Sri Rasiah HRB 247165 (AG München)
next parent reply other threads:[~2020-07-07 7:03 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20200702164411.108672-1-paul@darkrain42.org>
[not found] ` <878sfx6o64.fsf@suse.com>
[not found] ` <20200706192642.GA110607@haley.home.arpa>
2020-07-07 7:03 ` Aurélien Aptel [this message]
2020-07-07 13:05 ` lock checking issues (was: Re: [PATCH v3] cifs: Fix leak when handling lease break for cached root fid) Luc Van Oostenryck
2020-07-07 20:34 ` Paul Aurich
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=87tuyj6c3u.fsf@suse.com \
--to=aaptel@suse.com \
--cc=linux-cifs@vger.kernel.org \
--cc=linux-sparse@vger.kernel.org \
--cc=lsahlber@redhat.com \
--cc=paul@darkrain42.org \
--cc=sfrench@samba.org \
/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;
as well as URLs for NNTP newsgroup(s).