Hello,
On Mon, Mar 31, 2008 at 4:19 AM, TakashiYamamoto
<[EMAIL PROTECTED]> wrote:
> Hello,
> Thanks for your reply.
>
>
> > Why do you need to expose this Cell-specific feature into generic code?
> >
> > Can't you hide this inside the arch code by using the pfm_arch_*()
> callbacks?
>
> No.
> In this patch, I want to avoid setting TIF_PERFMON_CTXSW to the task flag
> when cell_spe_follow flag is set.
> I couln't hide this change in pfmon_arch_load_context(),
> because it is called before set_task_thread_flag().
>
Ok, I see your problem now.
We have 2 solutions:
- move pfm_arch_load_context() AFTER the TIF_PERFMON_CTXSW
- add a generic flag (context_flags) that pfm_arch_load_context()
can set to tell the generic code NOT to set TIF_PERFMON_CTXSW
I'd rather not move pfm_arch_load_context() because there would be
more code to undo if it ever fails.
The 2nd solution is not too pretty either because it relies on the
position of pfm_arch_load_context(). However, we could move
setting that flag to pfm_arch_create_context() given that you've added
the context-level flag for SPE_FOLLOW. Could you try that?
Thanks.
> Can I add a parameter which control to set TIF_PERFMON_CTXSW to the task flag
> to pfmon_arch_load_context()?
> OR
> Can I add pfmon_arch_set_task_thread_flag()?
>
>
> > There is an arch-specific portion to the context. You get to it with
> > ctx_arch =pfm_ctx_arch(ctx).
>
> OK,
> I can move cell_spe_follow flag to ctx_arch,
> If I can control to set TIF_PERFMON_CTXSW to the task flag from
> the arch specific code.
>
> Thanks.
> Takashi Yamamoto.
>
>
>
> >
> > 2008/3/28 TakashiYamamoto <[EMAIL PROTECTED]>:
> > > Hello,
> > > I updated the patch for the latest perfmon2 git tree.
> > >
> > > ---------------------------------------------------------------
> > >
> > > This patch adds cell_spe_follow flag to context flags.
> > > This flag is used for following the Cell SPE context switch.
> > >
> > > Signed-off-by: Takashi Yamamoto <[EMAIL PROTECTED]>
> > > ---
> > > arch/powerpc/perfmon/perfmon_cell.c | 2 ++
> > > include/asm-powerpc/perfmon.h | 8 ++++++++
> > > include/linux/perfmon_kern.h | 3 ++-
> > > perfmon/perfmon_attach.c | 26 ++++++++++++++++++++++----
> > >
> > >
> > > 4 files changed, 34 insertions(+), 5 deletions(-)
> > >
> > > --- a/arch/powerpc/perfmon/perfmon_cell.c
> > > +++ b/arch/powerpc/perfmon/perfmon_cell.c
> > > @@ -1220,6 +1220,8 @@ static void pfm_cell_get_ovfl_pmds(struc
> > > **/
> > > static int pfm_cell_create_context(struct pfm_context *ctx, u32
> ctx_flags)
> > > {
> > > + ctx->flags.cell_spe_follow =
> > > + (ctx_flags & PFM_FL_CELL_SPE_FOLLOW) ? 1:0;
> > > return 0;
> > > }
> > >
> > > --- a/include/asm-powerpc/perfmon.h
> > > +++ b/include/asm-powerpc/perfmon.h
> > > @@ -31,4 +31,12 @@
> > > #define PFM_ARCH_MAX_PMCS (256+64) /* 256 HW 64 SW */
> > > #define PFM_ARCH_MAX_PMDS (256+64) /* 256 HW 64 SW */
> > >
> > > +/*
> > > + * context flags (ctx_flags)
> > > + *
> > > + * bits[00-15]: generic flags
> > > + * bits[16-31]: arch-specific flags
> > > + */
> > > +#define PFM_FL_CELL_SPE_FOLLOW 0x10000
> > > +
> > > #endif /* _ASM_POWERPC_PERFMON_H_ */
> > > --- a/include/linux/perfmon_kern.h
> > > +++ b/include/linux/perfmon_kern.h
> > > @@ -118,7 +118,8 @@ struct pfm_context_flags {
> > > unsigned int can_restart:8; /* allowed to issue a
> PFM_RESTART */
> > > unsigned int reset_count:8; /* number of pending resets */
> > > unsigned int is_self:1; /* per-thread and self-montoring
> */
> > > - unsigned int reserved:5; /* for future use */
> > > + unsigned int cell_spe_follow:1; /* cell spe ctx follow */
> > > + unsigned int reserved:4; /* for future use */
> > > };
> > >
> > > /*
> > > --- a/perfmon/perfmon_attach.c
> > > +++ b/perfmon/perfmon_attach.c
> > > @@ -148,8 +148,17 @@ int __pfm_load_context(struct pfm_contex
> > >
> > > ctx->last_cpu = -1;
> > > set->priv_flags |= PFM_SETFL_PRIV_MOD_BOTH;
> > >
> > > - set_tsk_thread_flag(task, TIF_PERFMON_CTXSW);
> > > - PFM_DBG("[%d] set TIF", task->pid);
> > > + /*
> > > + * If cell_spe_follow is true, the task is not marked by
> > > + * TIF_PERFMON_CTXSW and pfm_ctxsw() is not called
> > > + * from the task scheduler.
> > > + * pfm_ctxsw() is called from SPU notifier in
> perfmon_cell.c
> > > + *
> > > + */
> > > + if (!ctx->flags.cell_spe_follow) {
> > > + set_tsk_thread_flag(task, TIF_PERFMON_CTXSW);
> > > + PFM_DBG("[%d] set TIF", task->pid);
> > > + }
> > >
> > > PFM_DBG("context loaded next ctxswin for [%d]",
> task->pid);
> > > } else {
> > > @@ -171,8 +180,17 @@ int __pfm_load_context(struct pfm_contex
> > >
> > >
> > >
> > > ctx->flags.is_self = 1;
> > >
> > > - set_tsk_thread_flag(task, TIF_PERFMON_CTXSW);
> > > - PFM_DBG("[%d] set TIF", task->pid);
> > > + /*
> > > + * If cell_spe_follow is true, the task is not marked by
> > > + * TIF_PERFMON_CTXSW and pfm_ctxsw() is not called
> > > + * from the task scheduler.
> > > + * pfm_ctxsw() is called from SPU notifier in
> perfmon_cell.c
> > > + *
> > > + */
> > > + if (!ctx->flags.cell_spe_follow) {
> > > + set_tsk_thread_flag(task, TIF_PERFMON_CTXSW);
> > > + PFM_DBG("[%d] set TIF", task->pid);
> > > + }
> > >
> > > /*
> > > * load PMD from set
> > >
> > >
> > >
> > > -------------------------------------------------------------------------
> > > Check out the new SourceForge.net Marketplace.
> > > It's the best place to buy or sell services for
> > > just about anything Open Source.
> > >
> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
> > > _______________________________________________
> > > perfmon2-devel mailing list
> > > [email protected]
> > > https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
> > >
> > >
>
>
-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
_______________________________________________
perfmon2-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
