Re: [net-next,1/3] netem: convert to qdisc_watchdog_schedule_ns

Re: [net-next,1/3] netem: convert to qdisc_watchdog_schedule_ns

[James Hogan]

[-- Attachment #1: Type: text/plain, Size: 984 bytes --]

On Tue, Nov 07, 2017 at 12:59:34PM -0800, Dave Taht wrote:
> diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c
> index db0228a..443a75d 100644
> --- a/net/sched/sch_netem.c
> +++ b/net/sched/sch_netem.c

...

> @@ -305,11 +305,11 @@ static bool loss_event(struct netem_sched_data *q)
>   * std deviation sigma.  Uses table lookup to approximate the desired
>   * distribution, and a uniformly-distributed pseudo-random source.
>   */
> -static psched_tdiff_t tabledist(psched_tdiff_t mu, psched_tdiff_t sigma,
> -				struct crndstate *state,
> -				const struct disttable *dist)
> +static s64 tabledist(s64 mu, s64 sigma,

sigma is used in a modulo operation in this function, which results in
this error on a bunch of MIPS configs once it is made 64-bits wide:

net/sched/sch_netem.o In function `tabledist':
net/sched/sch_netem.c:330: undefined reference to `__moddi3'

Should that code not be using <linux/math64.h>, i.e. div_s64_rem() now
that it is 64bit?

Thanks
James

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [net-next,1/3] netem: convert to qdisc_watchdog_schedule_ns

[James Hogan]

[-- Attachment #1: Type: text/plain, Size: 1270 bytes --]

On Tue, Nov 14, 2017 at 09:11:12PM +0000, James Hogan wrote:
> On Tue, Nov 07, 2017 at 12:59:34PM -0800, Dave Taht wrote:
> > diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c
> > index db0228a..443a75d 100644
> > --- a/net/sched/sch_netem.c
> > +++ b/net/sched/sch_netem.c
> 
> ...
> 
> > @@ -305,11 +305,11 @@ static bool loss_event(struct netem_sched_data *q)
> >   * std deviation sigma.  Uses table lookup to approximate the desired
> >   * distribution, and a uniformly-distributed pseudo-random source.
> >   */
> > -static psched_tdiff_t tabledist(psched_tdiff_t mu, psched_tdiff_t sigma,
> > -				struct crndstate *state,
> > -				const struct disttable *dist)
> > +static s64 tabledist(s64 mu, s64 sigma,
> 
> sigma is used in a modulo operation in this function, which results in
> this error on a bunch of MIPS configs once it is made 64-bits wide:
> 
> net/sched/sch_netem.o In function `tabledist':
> net/sched/sch_netem.c:330: undefined reference to `__moddi3'
> 
> Should that code not be using <linux/math64.h>, i.e. div_s64_rem() now
> that it is 64bit?

For the record, Dave has kindly pointed me at:
https://patchwork.ozlabs.org/project/netdev/list/?series=13554

which fixes the MIPS builds.

Cheers
James

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [net-next,1/3] netem: convert to qdisc_watchdog_schedule_ns

[Stephen Hemminger]

[-- Attachment #1: Type: text/plain, Size: 1249 bytes --]

On Tue, 14 Nov 2017 21:11:13 +0000
James Hogan <james.hogan@mips.com> wrote:

> On Tue, Nov 07, 2017 at 12:59:34PM -0800, Dave Taht wrote:
> > diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c
> > index db0228a..443a75d 100644
> > --- a/net/sched/sch_netem.c
> > +++ b/net/sched/sch_netem.c  
> 
> ...
> 
> > @@ -305,11 +305,11 @@ static bool loss_event(struct netem_sched_data *q)
> >   * std deviation sigma.  Uses table lookup to approximate the desired
> >   * distribution, and a uniformly-distributed pseudo-random source.
> >   */
> > -static psched_tdiff_t tabledist(psched_tdiff_t mu, psched_tdiff_t sigma,
> > -				struct crndstate *state,
> > -				const struct disttable *dist)
> > +static s64 tabledist(s64 mu, s64 sigma,  
> 
> sigma is used in a modulo operation in this function, which results in
> this error on a bunch of MIPS configs once it is made 64-bits wide:
> 
> net/sched/sch_netem.o In function `tabledist':
> net/sched/sch_netem.c:330: undefined reference to `__moddi3'
> 
> Should that code not be using <linux/math64.h>, i.e. div_s64_rem() now
> that it is 64bit?
> 
> Thanks
> James

Not really since random is only 32 bit, the sigma value being 64 bit makes
no sense really.

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]