| Message ID | 20221024195118.871907-1-code@tyhicks.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | clocksource/drivers/arm_arch_timer: Fix event stream param in Kconfig | expand |
Hi Tyler, On Mon, 24 Oct 2022 20:51:18 +0100, Tyler Hicks <code@tyhicks.com> wrote: > > From: "Tyler Hicks (Microsoft)" <code@tyhicks.com> > > Fix the event stream timer command line parameter name that's documented > in the Kconfig description for CONFIG_ARM_ARCH_TIMER_EVTSTREAM. It > didn't match the command line parameter name that's actually honored in > the source code. > > Reported-by: Vijay Balakrishna <vijayb@linux.microsoft.com> > Fixes: 46fd5c6b3059 ("clocksource/drivers/arm_arch_timer: Control the evtstrm via the cmdline") > Cc: stable@vger.kernel.org No, this really doesn't deserve a Cc: stable. This bit may be wrong, but we have the correct information in the kernel-parameters.txt file, which is authoritative AFAIC. > Signed-off-by: Tyler Hicks (Microsoft) <code@tyhicks.com> > --- > drivers/clocksource/Kconfig | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/clocksource/Kconfig b/drivers/clocksource/Kconfig > index 4469e7f555e9..dcc274460973 100644 > --- a/drivers/clocksource/Kconfig > +++ b/drivers/clocksource/Kconfig > @@ -315,7 +315,7 @@ config ARM_ARCH_TIMER_EVTSTREAM > executing the wfe instruction at a frequency represented as a > power-of-2 divisor of the clock rate. The behaviour can also be > overridden on the command line using the > - clocksource.arm_arch_timer.evtstream parameter. > + clocksource.arm_arch_timer.evtstrm parameter. > The main use of the event stream is wfe-based timeouts of userspace > locking implementations. It might also be useful for imposing timeout > on wfe to safeguard against any programming errors in case an expected With this tag dropped, Acked-by: Marc Zyngier <maz@kernel.org> M.
On 2022-10-25 11:24:55, Marc Zyngier wrote: > Hi Tyler, > > On Mon, 24 Oct 2022 20:51:18 +0100, > Tyler Hicks <code@tyhicks.com> wrote: > > > > From: "Tyler Hicks (Microsoft)" <code@tyhicks.com> > > > > Fix the event stream timer command line parameter name that's documented > > in the Kconfig description for CONFIG_ARM_ARCH_TIMER_EVTSTREAM. It > > didn't match the command line parameter name that's actually honored in > > the source code. > > > > Reported-by: Vijay Balakrishna <vijayb@linux.microsoft.com> > > Fixes: 46fd5c6b3059 ("clocksource/drivers/arm_arch_timer: Control the evtstrm via the cmdline") > > Cc: stable@vger.kernel.org > > No, this really doesn't deserve a Cc: stable. This bit may be wrong, > but we have the correct information in the kernel-parameters.txt file, > which is authoritative AFAIC. Thanks for the review! I added the stable tag because it caused an actual waste of time in an investigation. Upon discovering this option, I didn't notice the subtle difference in cmdline parameter name from the Kconfig, to the kernel-parameters.txt file and the code. I ran long-running stress tests with the incorrect parameter name copied and pasted from the Kconfig. This error resulted in a bad data point in the investigation and it took a second set of eyes (Vijay's) to uncover my mistake days later when we were trying to reconcile the bad data point with other good data points. It is no big deal if it is merged with the tag dropped but I wanted to provide my reasoning of why I personally felt this was a more than just a comment typo fix or something along those lines. Thanks, again! Tyler > > > Signed-off-by: Tyler Hicks (Microsoft) <code@tyhicks.com> > > --- > > drivers/clocksource/Kconfig | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/clocksource/Kconfig b/drivers/clocksource/Kconfig > > index 4469e7f555e9..dcc274460973 100644 > > --- a/drivers/clocksource/Kconfig > > +++ b/drivers/clocksource/Kconfig > > @@ -315,7 +315,7 @@ config ARM_ARCH_TIMER_EVTSTREAM > > executing the wfe instruction at a frequency represented as a > > power-of-2 divisor of the clock rate. The behaviour can also be > > overridden on the command line using the > > - clocksource.arm_arch_timer.evtstream parameter. > > + clocksource.arm_arch_timer.evtstrm parameter. > > The main use of the event stream is wfe-based timeouts of userspace > > locking implementations. It might also be useful for imposing timeout > > on wfe to safeguard against any programming errors in case an expected > > With this tag dropped, > > Acked-by: Marc Zyngier <maz@kernel.org> > > M. > > -- > Without deviation from the norm, progress is not possible.
Hi Tyler, On Tue, 25 Oct 2022 15:18:54 +0100, "Tyler Hicks (Microsoft)" <code@tyhicks.com> wrote: > > On 2022-10-25 11:24:55, Marc Zyngier wrote: > > Hi Tyler, > > > > On Mon, 24 Oct 2022 20:51:18 +0100, > > Tyler Hicks <code@tyhicks.com> wrote: > > > > > > From: "Tyler Hicks (Microsoft)" <code@tyhicks.com> > > > > > > Fix the event stream timer command line parameter name that's documented > > > in the Kconfig description for CONFIG_ARM_ARCH_TIMER_EVTSTREAM. It > > > didn't match the command line parameter name that's actually honored in > > > the source code. > > > > > > Reported-by: Vijay Balakrishna <vijayb@linux.microsoft.com> > > > Fixes: 46fd5c6b3059 ("clocksource/drivers/arm_arch_timer: Control the evtstrm via the cmdline") > > > Cc: stable@vger.kernel.org > > > > No, this really doesn't deserve a Cc: stable. This bit may be wrong, > > but we have the correct information in the kernel-parameters.txt file, > > which is authoritative AFAIC. > > Thanks for the review! > > I added the stable tag because it caused an actual waste of time in an > investigation. Upon discovering this option, I didn't notice the subtle > difference in cmdline parameter name from the Kconfig, to the > kernel-parameters.txt file and the code. I ran long-running stress tests > with the incorrect parameter name copied and pasted from the Kconfig. > This error resulted in a bad data point in the investigation and it took > a second set of eyes (Vijay's) to uncover my mistake days later when we > were trying to reconcile the bad data point with other good data points. I appreciate this has caused a certain waste of time, but the rules for stable inclusion are pretty clear: it has to be a runtime bug. Documentation bugs, as irritating as they are, do not fall into this category (it really *is* a typo). Thanks, M.
diff --git a/drivers/clocksource/Kconfig b/drivers/clocksource/Kconfig index 4469e7f555e9..dcc274460973 100644 --- a/drivers/clocksource/Kconfig +++ b/drivers/clocksource/Kconfig @@ -315,7 +315,7 @@ config ARM_ARCH_TIMER_EVTSTREAM executing the wfe instruction at a frequency represented as a power-of-2 divisor of the clock rate. The behaviour can also be overridden on the command line using the - clocksource.arm_arch_timer.evtstream parameter. + clocksource.arm_arch_timer.evtstrm parameter. The main use of the event stream is wfe-based timeouts of userspace locking implementations. It might also be useful for imposing timeout on wfe to safeguard against any programming errors in case an expected