On Thu, May 29, 2014 at 09:02:36AM +0900, Namhyung Kim wrote:
> On Tue, 27 May 2014 09:37:38 +0200, Jiri Olsa wrote:
> > On Tue, May 27, 2014 at 10:05:28AM +0900, Namhyung Kim wrote:
> >> Hi Jiri,
> >>
> >> On Thu, 15 May 2014 19:23:27 +0200, Jiri Olsa wrote:
> >>
> >> [SNIP]
> >> > +static void
On Tue, 27 May 2014 09:37:38 +0200, Jiri Olsa wrote:
> On Tue, May 27, 2014 at 10:05:28AM +0900, Namhyung Kim wrote:
>> Hi Jiri,
>>
>> On Thu, 15 May 2014 19:23:27 +0200, Jiri Olsa wrote:
>>
>> [SNIP]
>> > +static void data_close(void)
>> > +{
>> > + bool cache_fd = may_cache_fd();
>> > +
>> > +
On Tue, May 27, 2014 at 10:05:28AM +0900, Namhyung Kim wrote:
> Hi Jiri,
>
> On Thu, 15 May 2014 19:23:27 +0200, Jiri Olsa wrote:
>
> [SNIP]
> > +static void data_close(void)
> > +{
> > + bool cache_fd = may_cache_fd();
> > +
> > + if (!cache_fd)
> > + close_first_dso();
> > +}
>
>
Hi Jiri,
On Thu, 15 May 2014 19:23:27 +0200, Jiri Olsa wrote:
[SNIP]
> +static void data_close(void)
> +{
> + bool cache_fd = may_cache_fd();
> +
> + if (!cache_fd)
> + close_first_dso();
> +}
Why do you do this at close()? As long as there's no attempt to open a
new file, w
Changing the dso__data_close to keep the dso data file
descriptor open if possible.
We keep dsos data file descriptors open until their
count reaches the half of the current fd open limit
(RLIMIT_NOFILE). In this case we close file descriptor
of the first opened dso object.
We've got speed up in
5 matches
Mail list logo