On Fri, Aug 29, 2014 at 4:48 PM, Yann Ylavic wrote:
> On Fri, Aug 29, 2014 at 4:28 PM, Ruediger Pluem wrote:
>> Is this correct? apr_atomic_add32 expects an uint as 2nd argument. I just
>> experienced a situation where all workers are
>> idle, but queue_info->idlers < zero_pt.
>
> The final code
On Fri, Aug 29, 2014 at 4:28 PM, Ruediger Pluem wrote:
>
>
> j...@apache.org wrote:
>> Author: jim
>> Date: Mon Nov 25 13:59:06 2013
>> New Revision: 1545286
>>
>> URL: http://svn.apache.org/r1545286
>> Log:
>> Use a normalized offset point for idlers... still need to worry
>> that atomics work as
6 - in
> /httpd/httpd/trunk/server/mpm/event: event.c fdqueue.c
>
> This operation is checked in event... see the diffs for event.c
> in that commit.
>
> On Aug 29, 2014, at 10:28 AM, Ruediger Pluem wrote:
>
> >
> >
> > j...@apache.org wrote:
> &g
This operation is checked in event... see the diffs for event.c
in that commit.
On Aug 29, 2014, at 10:28 AM, Ruediger Pluem wrote:
>
>
> j...@apache.org wrote:
>> Author: jim
>> Date: Mon Nov 25 13:59:06 2013
>> New Revision: 1545286
>>
>> URL: http://svn.apache.org/r1545286
>> Log:
>> Use a
j...@apache.org wrote:
> Author: jim
> Date: Mon Nov 25 13:59:06 2013
> New Revision: 1545286
>
> URL: http://svn.apache.org/r1545286
> Log:
> Use a normalized offset point for idlers... still need to worry
> that atomics work as "expected", in this case that a add32 of a -1
> is the "same" as d
On Mon, Nov 25, 2013 at 7:40 PM, Jim Jagielski wrote:
>
> On Nov 25, 2013, at 1:09 PM, Jim Jagielski wrote:
>
> >
> > On Nov 25, 2013, at 11:53 AM, Yann Ylavic wrote:
> >
> >> On Mon, Nov 25, 2013 at 4:53 PM, Jim Jagielski wrote:
> >>
> >> On Nov 25, 2013, at 10:24 AM, Yann Ylavic wrote:
> >>
On Mon, Nov 25, 2013 at 7:10 PM, Jim Jagielski wrote:
>
> On Nov 25, 2013, at 11:53 AM, Yann Ylavic wrote:
> >
> > Indeed, that was not consistent!
> > This is a second fix then, the check was doubly broken...
> >
>
> I'm not sure...
>
Yes, APR_EAGAIN is "all workers busy", so the <= 1 was real
Upon review, using unsigned makes a lot of sense, so I'll
start the adjustment from apr_int32_t and simple ints
to apr_uint32_t.
On Nov 25, 2013, at 1:40 PM, Jim Jagielski wrote:
>
> On Nov 25, 2013, at 1:09 PM, Jim Jagielski wrote:
>
>>
>> On Nov 25, 2013, at 11:53 AM, Yann Ylavic wrote:
>
On Nov 25, 2013, at 1:09 PM, Jim Jagielski wrote:
>
> On Nov 25, 2013, at 11:53 AM, Yann Ylavic wrote:
>
>> On Mon, Nov 25, 2013 at 4:53 PM, Jim Jagielski wrote:
>>
>> On Nov 25, 2013, at 10:24 AM, Yann Ylavic wrote:
>>
>>>
>>> As Jeff said in the other thread, I think the test here shou
On Nov 25, 2013, at 11:53 AM, Yann Ylavic wrote:
>
> Indeed, that was not consistent!
> This is a second fix then, the check was doubly broken...
>
I'm not sure...
On Nov 25, 2013, at 11:53 AM, Yann Ylavic wrote:
> On Mon, Nov 25, 2013 at 4:53 PM, Jim Jagielski wrote:
>
> On Nov 25, 2013, at 10:24 AM, Yann Ylavic wrote:
>
> >
> > As Jeff said in the other thread, I think the test here should be :
> >if (prev_idlers <= 1)
> > that's because dec32 wa
On Mon, Nov 25, 2013 at 4:24 PM, Yann Ylavic wrote:
>
> (Note the "int prev_idlers" => "apr_int32_t prev_idlers" too in the patch,
> should int be larger than int32_t, eg. with 64bits int, prev_idlers would
> not become negative is this case...).
>
Just to illustrate, suppose sizeof(int) == size
On Mon, Nov 25, 2013 at 4:53 PM, Jim Jagielski wrote:
>
> On Nov 25, 2013, at 10:24 AM, Yann Ylavic wrote:
>
> >
> > As Jeff said in the other thread, I think the test here should be :
> >if (prev_idlers <= 1)
> > that's because dec32 was returning the new value while add32 now returns
> the
On Nov 25, 2013, at 10:24 AM, Yann Ylavic wrote:
>
> As Jeff said in the other thread, I think the test here should be :
>if (prev_idlers <= 1)
> that's because dec32 was returning the new value while add32 now returns the
> old one (fetch_and_sub vs sub_and_fetch).
>
We aren't being con
On Mon, Nov 25, 2013 at 4:24 PM, Yann Ylavic wrote:
>
> Index: server/mpm/event/fdqueue.c
> ===
> --- server/mpm/event/fdqueue.c(revision 1545301)
> +++ server/mpm/event/fdqueue.c(working copy)
> @@ -17,7 +17,7 @@
> #include
On Mon, Nov 25, 2013 at 2:59 PM, wrote:
> Author: jim
> Date: Mon Nov 25 13:59:06 2013
> New Revision: 1545286
>
> URL: http://svn.apache.org/r1545286
> Log:
> Use a normalized offset point for idlers... still need to worry
> that atomics work as "expected", in this case that a add32 of a -1
> is
16 matches
Mail list logo