Re: [PATCH 1/3] make sparse keep its promise about context tracking

[Josh Triplett <josh@freedesktop.org>]

[-- Attachment #1: Type: text/plain, Size: 1011 bytes --]

One comment on your patch description, by the way:

Johannes Berg wrote:
> However, the kernel use of __attribute__((context(...)) is actually
> wrong, the kernel often does things like:
> 
>     static void *dev_mc_seq_start(struct seq_file *seq, loff_t * pos)
>             __acquires(dev_base_lock)
>     {
>             [...]
>             read_lock(&dev_base_lock);
>             [...]
>     }
> 
> rather than
> 
>     static void *dev_mc_seq_start(struct seq_file *seq, loff_t * pos)
>             __acquires(dev_base_lock)
>     {
>             [...]
>             __acquire__(dev_base_lock);
>             read_lock(&dev_base_lock);
>             [...]
>     }

I don't understand why you count this as wrong.  read_lock should
handle acquiring the context itself, either via an __acquires
annotation if written as a function, or via an __acquire__ statement
if written as a macro.  Callers of read_lock shouldn't need to
explicitly call __acquire__.

- Josh Triplett


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 252 bytes --]