On Wed, 30 May 2012 16:49:02 +0200
Eric Dumazet wrote:
> On Wed, 2012-05-30 at 20:29 +0900, Hiroaki SHIMODA wrote:
> > On Wed, 30 May 2012 13:08:27 +0200
> > Eric Dumazet wrote:
> >
> > > On Wed, 2012-05-30 at 19:43 +0900, Hiroaki SHIMODA wrote:
> > >
> > > > While examining ping problem, belo
The problem is that is only happens on servers getting a reasonable (>
~5000 pps and greater than 50Mbit). So, idle servers’ error counters do not
increase.
Nonetheless, I still tried a sniff on a server getting at least 1
rx_short_length_errors per second. I could not find any anomalies in the
On 05/30/2012 07:49 AM, Eric Dumazet wrote:
> On Wed, 2012-05-30 at 20:29 +0900, Hiroaki SHIMODA wrote:
>> On Wed, 30 May 2012 13:08:27 +0200
>> Eric Dumazet wrote:
>>
>>> On Wed, 2012-05-30 at 19:43 +0900, Hiroaki SHIMODA wrote:
>>>
While examining ping problem, below pattern is often observ
On Wed, 2012-05-30 at 16:42 +0200, Eric Dumazet wrote:
> On Wed, 2012-05-30 at 07:09 -0700, Joe Perches wrote:
>
> > Whatever evals A and B once would be better.
>
> Why do you believe they could be evaluated several time ?
>
> #define POSDIFF(A, B) max_t(int, (A) - (B), 0)
I don't and didn't s
On Wed, 2012-05-30 at 20:29 +0900, Hiroaki SHIMODA wrote:
> On Wed, 30 May 2012 13:08:27 +0200
> Eric Dumazet wrote:
>
> > On Wed, 2012-05-30 at 19:43 +0900, Hiroaki SHIMODA wrote:
> >
> > > While examining ping problem, below pattern is often observed.
> > >
> > >
On Wed, 2012-05-30 at 07:09 -0700, Joe Perches wrote:
> Whatever evals A and B once would be better.
Why do you believe they could be evaluated several time ?
#define POSDIFF(A, B) max_t(int, (A) - (B), 0)
(A) - (B) is done once, or its a huge max_t() bug.
By the way, we handle 32bit values he
On Wed, 2012-05-30 at 13:59 +0200, Eric Dumazet wrote:
> On Wed, 2012-05-30 at 04:20 -0700, Joe Perches wrote:
> > On Wed, 2012-05-30 at 13:08 +0200, Eric Dumazet wrote:
> > > Maybe we should change all POSDIFF(), not selected ones.
> > > #define POSDIFF(A, B) ((int)((A) - (B)) > 0 ? (A) - (B) : 0)
On Wed, 2012-05-30 at 04:20 -0700, Joe Perches wrote:
> On Wed, 2012-05-30 at 13:08 +0200, Eric Dumazet wrote:
> > Maybe we should change all POSDIFF(), not selected ones.
> > #define POSDIFF(A, B) ((int)((A) - (B)) > 0 ? (A) - (B) : 0)
>
> maybe use an eval once statement expression macro
> ({
>
On Wed, 30 May 2012 13:08:27 +0200
Eric Dumazet wrote:
> On Wed, 2012-05-30 at 19:43 +0900, Hiroaki SHIMODA wrote:
>
> > While examining ping problem, below pattern is often observed.
> >
> >TIME
> >dql_queued() dql_completed()
On Wed, 2012-05-30 at 13:08 +0200, Eric Dumazet wrote:
> Maybe we should change all POSDIFF(), not selected ones.
> #define POSDIFF(A, B) ((int)((A) - (B)) > 0 ? (A) - (B) : 0)
maybe use an eval once statement expression macro
({
typeof (A) _a = (A);
typeof (B) _b = (B);
((
> > + num_queued = dql->num_queued;
>
>
> I suggest :
>
> num_queued = ACCESS_ONCE(dql->num_queued);
>
> Or else compiler is free to do whatever he wants.
Or make the structure member volatile, then the
compiler can only read it once.
Probably worth while if the value is expect
On Wed, 2012-05-30 at 13:04 +0200, Eric Dumazet wrote:
> Please don't use lazy volatile.
>
> Unless you really want Linus flames (and ours)
>
> ACCESS_ONCE() is much cleaner.
>
http://yarchive.net/comp/linux/ACCESS_ONCE.html
--
On Wed, 2012-05-30 at 19:43 +0900, Hiroaki SHIMODA wrote:
> While examining ping problem, below pattern is often observed.
>
>TIME
>dql_queued() dql_completed() |
> a) initial state |
>
On Wed, 2012-05-30 at 11:52 +0100, David Laight wrote:
> > > + num_queued = dql->num_queued;
> >
> >
> > I suggest :
> >
> > num_queued = ACCESS_ONCE(dql->num_queued);
> >
> > Or else compiler is free to do whatever he wants.
>
> Or make the structure member volatile, then the
> compil
On Wed, 30 May 2012 10:40:31 +0200
Eric Dumazet wrote:
> On Wed, 2012-05-30 at 09:06 +0900, Hiroaki SHIMODA wrote:
> > While reading the bql code, I have some questions.
> >
> > 1) dql_completed() and dql_queued() can be called concurrently,
> >so dql->num_queued could change while processin
On Wed, 2012-05-30 at 09:06 +0900, Hiroaki SHIMODA wrote:
> While reading the bql code, I have some questions.
>
> 1) dql_completed() and dql_queued() can be called concurrently,
>so dql->num_queued could change while processing
>dql_completed().
>Is it intentional to refer num_queued
16 matches
Mail list logo
