Re: A question about sparse: how to use __acquires() and __releases() correctly ?

[Eric Dumazet <dada1@cosmosbay.com>]

Bart Van Assche a écrit :
> [ping]
> 
> Is there anyone who can help me with the question below ?
> 
> Thanks,
> 
> Bart.
> 
> On Tue, Dec 2, 2008 at 8:59 PM, Bart Van Assche
> <bart.vanassche@gmail.com> wrote:
>> Hello,
>>
>> I'm helping to prepare the SCST source code for inclusion in the Linux
>> kernel by a.o. cleaning up sparse warnings. Although most of the SCST
>> source code has been annotated by this time it's still not clear to me
>> how to use __acquires() and __releases() correctly.
>>
>> I will illustrate my questions via the following code from
>> net/core/dev.c, Linux kernel version 2.6.27.7:
>>
>> void dev_seq_stop(struct seq_file *seq, void *v)
>>        __releases(dev_base_lock)
>> {
>>        read_unlock(&dev_base_lock);
>> }
>>
>> The command "make C=2 M=net/core" produces the following output for
>> the above function (using a sparse binary built from the sparse git
>> repository, last updated on August 26, 2008):
>>
>> net/core/dev.c:2579:2: warning: context problem in 'dev_seq_stop':
>> '_read_unlock' expected different context
>> net/core/dev.c:2579:2:    context 'lock': wanted >= 1, got 0
>>
>> My questions are as follows:
>> * Which argument type should be passed to __releases() -- a pointer to
>> a lock structure or the lock strucure itself ? In the header file
>> include/linux/spinlock_api_smp.h a pointer is passed to __acquires()
>> and __releases(), while other code (like the above) passes the lock
>> structure itself to the __acquires() and __releases() annotations.
>> * If the __releases() annotation is used correctly in net/core/dev.c,
>> why does sparse complain about a context problem ?
>>
>> Please keep me in CC -- I'm not subscribed to the LKML.
>>
>> Bart.
>>

I added __releases() annotation in dev_seq_stop() in January 1st

^1da177e (Linus Torvalds                 2005-04-16 15:20:36 -0700 2617) }
^1da177e (Linus Torvalds                 2005-04-16 15:20:36 -0700 2618)
^1da177e (Linus Torvalds                 2005-04-16 15:20:36 -0700 2619) void dev_seq_stop(struct seq_file *seq, void *v)
9a429c49 (Eric Dumazet                   2008-01-01 21:58:02 -0800 2620)        __releases(dev_base_lock)
^1da177e (Linus Torvalds                 2005-04-16 15:20:36 -0700 2621) {
^1da177e (Linus Torvalds                 2005-04-16 15:20:36 -0700 2622)        read_unlock(&dev_base_lock);
^1da177e (Linus Torvalds                 2005-04-16 15:20:36 -0700 2623) }


The sparse version I have here doesnt complain on dev_seq_stop() in net/core/dev.c

  CHECK   net/core/dev.c
net/core/dev.c:2060:29: warning: symbol 'br_fdb_get_hook' was not declared. Should it be static?
net/core/dev.c:2062:6: warning: symbol 'br_fdb_put_hook' was not declared. Should it be static?
net/core/dev.c:1889:4: warning: context imbalance in 'dev_queue_xmit' - different lock contexts for basic block
net/core/dev.c:2022:3: warning: context imbalance in 'net_tx_action' - different lock contexts for basic block
include/linux/netpoll.h:94:2: warning: context imbalance in 'net_rx_action' - different lock contexts for basic block
  CC      net/core/dev.o

# sparse --version
# sparse -V
# type sparse
sparse is hashed (/root/bin/sparse)
# ls -l /root/bin/sparse
-rwxr-xr-x  1 root root 947324 déc 17  2007 /root/bin/sparse
# cd /usr/src/sparse
# git log | head
commit a02aeb329d5a8f9047c0b75b7e7f64ee2db3ffcf
Author: Josh Triplett <josh@freedesktop.org>
Date:   Tue Nov 13 04:15:13 2007 -0800

    Makefile: VERSION=0.4.1

    Signed-off-by: Josh Triplett <josh@freedesktop.org>