RE: [PATCH] power: print function name of callbacks

From: Brown, Len
Date: Tue Sep 22 2015 - 13:33:30 EST


have you used analyze_suspend?
It used to parse this output, but that was abandoned
when it cut over to using ftrace directly.

https://01.org/suspendresume

cheers,
-Len


> -----Original Message-----
> From: Douglas Anderson [mailto:dianders@xxxxxxxxxxxx]
> Sent: Tuesday, September 22, 2015 1:27 PM
> To: rjw@xxxxxxxxxxxxx
> Cc: Dmitry Torokhov; Douglas Anderson; pavel@xxxxxx; Brown, Len;
> gregkh@xxxxxxxxxxxxxxxxxxx; linux-pm@xxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx
> Subject: [PATCH] power: print function name of callbacks
>
> The printouts writen to the logs by suspend can be a bit opaque: it can
> be hard to track them down to the actual function called. You might
> see:
>
> calling rfkill1+ @ 19473, parent: phy0
> call rfkill1+ returned 0 after 1 usecs
> calling phy0+ @ 19473, parent: mmc2:0001:1
> call phy0+ returned 0 after 19 usecs
>
> It's a bit hard to know what's actually happening. Instead, it's nice
> to see:
>
> calling rfkill1+ @ 15793, parent: phy0, cb: rfkill_suspend
> call rfkill1+ returned 0 after 1 usecs
> calling phy0+ @ 15793, parent: mmc2:0001:1, cb: wiphy_suspend
> [cfg80211]
> call phy0+ returned 0 after 7 usecs
>
> That makes it very obvious what's going on. It also has the nice side
> effect of making the suspend/resume spew a little more obvious, since
> many resume functions have the word "resume" in the name:
>
> calling phy0+ @ 15793, parent: mmc2:0001:1, cb: wiphy_resume [cfg80211]
> call phy0+ returned 0 after 12 usecs
> calling rfkill1+ @ 15793, parent: phy0, cb: rfkill_resume
> call rfkill1+ returned 0 after 1 usecs
>
> Signed-off-by: Douglas Anderson <dianders@xxxxxxxxxxxx>
> ---
> drivers/base/power/main.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c
> index 1710c26..e2b1f14 100644
> --- a/drivers/base/power/main.c
> +++ b/drivers/base/power/main.c
> @@ -188,14 +188,14 @@ void device_pm_move_last(struct device *dev)
> list_move_tail(&dev->power.entry, &dpm_list);
> }
>
> -static ktime_t initcall_debug_start(struct device *dev)
> +static ktime_t initcall_debug_start(struct device *dev, void *cb)
> {
> ktime_t calltime = ktime_set(0, 0);
>
> if (pm_print_times_enabled) {
> - pr_info("calling %s+ @ %i, parent: %s\n",
> + pr_info("calling %s+ @ %i, parent: %s, cb: %pf\n",
> dev_name(dev), task_pid_nr(current),
> - dev->parent ? dev_name(dev->parent) : "none");
> + dev->parent ? dev_name(dev->parent) : "none", cb);
> calltime = ktime_get();
> }
>
> @@ -382,7 +382,7 @@ static int dpm_run_callback(pm_callback_t cb, struct
> device *dev,
> if (!cb)
> return 0;
>
> - calltime = initcall_debug_start(dev);
> + calltime = initcall_debug_start(dev, cb);
>
> pm_dev_dbg(dev, state, info);
> trace_device_pm_callback_start(dev, info, state.event);
> @@ -1324,7 +1324,7 @@ static int legacy_suspend(struct device *dev,
> pm_message_t state,
> int error;
> ktime_t calltime;
>
> - calltime = initcall_debug_start(dev);
> + calltime = initcall_debug_start(dev, cb);
>
> trace_device_pm_callback_start(dev, info, state.event);
> error = cb(dev, state);
> --
> 2.6.0.rc0.131.gf624c3d

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/