Re: [tip: timers/clocksource] clocksource/drivers/exynos_mct: Don't register as a sched_clock on arm64
From: Ingo Molnar
Date: Fri Jul 25 2025 - 06:37:37 EST
* Daniel Lezcano <daniel.lezcano@xxxxxxxxxx> wrote:
> On 24/07/2025 07:16, Ingo Molnar wrote:
> >
> > * tip-bot2 for Will McVicker <tip-bot2@xxxxxxxxxxxxx> wrote:
> >
> > > The following commit has been merged into the timers/clocksource branch of tip:
> > >
> > > Commit-ID: 394b981382e6198363cf513f6eb6be4c55b22e44
> > > Gitweb: https://git.kernel.org/tip/394b981382e6198363cf513f6eb6be4c55b22e44
> > > Author: Will McVicker <willmcvicker@xxxxxxxxxx>
> > > AuthorDate: Fri, 20 Jun 2025 11:17:05 -07:00
> > > Committer: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
> > > CommitterDate: Tue, 15 Jul 2025 13:00:50 +02:00
> > >
> > > clocksource/drivers/exynos_mct: Don't register as a sched_clock on arm64
> > >
> > > The MCT register is unfortunately very slow to access, but importantly
> > > does not halt in the c2 idle state. So for ARM64, we can improve
> > > performance by not registering the MCT for sched_clock, allowing the
> > > system to use the faster ARM architected timer for sched_clock instead.
> > >
> > > The MCT is still registered as a clocksource, and a clockevent in order
> > > to be a wakeup source for the arch_timer to exit the "c2" idle state.
> > >
> > > Since ARM32 SoCs don't have an architected timer, the MCT must continue
> > > to be used for sched_clock. Detailed discussion on this topic can be
> > > found at [1].
> > >
> > > [1] https://lore.kernel.org/linux-samsung-soc/1400188079-21832-1-git-send-email-chirantan@xxxxxxxxxxxx/
> > >
> > > [Original commit from https://android.googlesource.com/kernel/gs/+/630817f7080e92c5e0216095ff52f6eb8dd00727
> > >
> > > Signed-off-by: Donghoon Yu <hoony.yu@xxxxxxxxxxx>
> > > Signed-off-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
> > > Reviewed-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
> > > Acked-by: John Stultz <jstultz@xxxxxxxxxx>
> > > Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
> > > Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
> > > Link: https://lore.kernel.org/r/20250620181719.1399856-3-willmcvicker@xxxxxxxxxx
> > > Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
> >
> > The whole SOB chain of this commit is messy and has several serious
> > problems:
> >
> > 1)
> >
> > This commit has misattributed authorship: the first SOB is:
> >
> > Signed-off-by: Donghoon Yu <hoony.yu@xxxxxxxxxxx>
> >
> > but the Author field is not Donghoon Yu:
> >
> > Author: Will McVicker <willmcvicker@xxxxxxxxxx>
>
> Yes, you are right. I should have pay more attention to author / sob, thanks
> for spotting it.
>
> > 2)
> >
> > The Reviewed-by tag is misapplied:
> >
> > > Signed-off-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
> > > Reviewed-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
> >
> > When someone passes along a patch, it's implicit that they have
> > reviewed it.
>
> Well my understanding of the SOB chain for these is the Signed-off-by from
> Youngmin is in the delivery path because it went first to the AOSP, then
> carried on to Linux by Will. Then Reviewed-by Youngmin letting us know the
> port from AOSP to Linux is ok.
>
> > 3)
> >
> > There's also a stray Tested-by tag by one of the SOB entries:
> >
> > > Signed-off-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
> > > Reviewed-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
> > > Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
> >
> > When someone passes along a patch, it's implicit that they not only
> > have reviewed the patch, but have also tested it to a certain extent
>
> In this specific case where the original commit is from AOSP, this chain
> seems to make sense. Souns like:
>
> "I was in the original commit delivery path"
> "I reviewed this patch carried to Linux"
> "I tested it on Linux"
Yeah, so then this should be documented by adding a comment to the tag
itself:
Signed-off-by: Donghoon Yu <hoony.yu@xxxxxxxxxxx>
Signed-off-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx> # AOSP -> Linux port
Reviewed-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx> # AOSP -> Linux port
Otherwise it's just confusing as to why there's duplicate SOB and
Reviewed-by entries.
But as long as the porting was basically just a cherry-pick, these
extra tags are probably superfluous. If there was a conflict resolved
by one of the maintainers along the SOB chain, that should be marked
explicitly, which I see was already done in some cases:
[ dlezcano : Fixed conflict with 20250614175556.922159-2-linux@xxxxxxxxxxxx ]
> > ...
> >
> > 4)
> >
> > Why is the 'Link' tag just in the middle of the SOB chain, instead at the end of it?
>
> I don't know. Link must be at the end It is stated somewhere in the
> documentation?
>
> I use git b4 -s <msg-id> and the tool adds the Link then my sign off.
Yeah, so using tools and not looking at the end result will often just
create a random tag order that looks messy.
On preferred tag ordering, see:
Documentation/process/maintainer-tip.rst
Ordering of commit tags
^^^^^^^^^^^^^^^^^^^^^^^
...
'Link' is at the end of the list of tags.
There's some logic to the -tip tag ordering (more important tags go
before less important tags), but it's mostly just an arbitrary order
that we try to stick to within -tip.
> > Presumably this is the proper SOB chain:
> >
> > > Author: Donghoon Yu <hoony.yu@xxxxxxxxxxx>
> >
> > > Signed-off-by: Donghoon Yu <hoony.yu@xxxxxxxxxxx>
> > > Signed-off-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
> > > Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
> > > Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
> > > Acked-by: John Stultz <jstultz@xxxxxxxxxx>
> > > Link: https://lore.kernel.org/r/20250620181719.1399856-3-willmcvicker@xxxxxxxxxx
> >
> > Correct?
>
>
So I got no answer for this question, but I suppose my assumption is
correct - so I've rebased the tip:timers/clocksource commits to fix the
misattribution and a number of other problems, and also fixed various
typos, spelling mistakes and inconsistencies in the changelogs while at
it. Let me know if I got something wrong.
I've attached a delta-patch of the changelog changes below - note that
I skipped the commit IDs to make the diff easier to read.
Thanks,
Ingo
===================>
--- 2025-07-25 12:15:26.024284067 +0200
+++ 2025-07-25 12:15:18.761435799 +0200
@@ -10,6 +10,7 @@ Date: Tue Jul 15 14:18:33 2025 +0200
the module loading.
Fix this by adding the __init_or_module annotation for the functions:
+
- mct_init_dt()
- mct_init_spi()
- mct_init_dt()
@@ -17,9 +18,10 @@ Date: Tue Jul 15 14:18:33 2025 +0200
Compiled on ARM32 + MODULES=no, ARM64 + MODULES=yes, ARM64 +
MODULES=no
- Link: https://lore.kernel.org/r/20250715121834.2059191-1-daniel.lezcano@xxxxxxxxxx
- Reviewed-by: Will McVicker <willmcvicker@xxxxxxxxxx>
Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
+ Reviewed-by: Will McVicker <willmcvicker@xxxxxxxxxx>
+ Link: https://lore.kernel.org/r/20250715121834.2059191-1-daniel.lezcano@xxxxxxxxxx
Author: AngeloGioacchino Del Regno <angelogioacchino.delregno@xxxxxxxxxxxxx>
Date: Wed Jun 11 13:07:58 2025 +0200
@@ -30,11 +32,12 @@ Date: Wed Jun 11 13:07:58 2025 +0200
with MT6765.
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@xxxxxxxxxxxxx>
- Acked-by: Rob Herring (Arm) <robh@xxxxxxxxxx>
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Reviewed-by: Nícolas F. R. A. Prado <nfraprado@xxxxxxxxxxxxx>
+ Acked-by: Rob Herring (Arm) <robh@xxxxxxxxxx>
Acked-by: Conor Dooley <conor.dooley@xxxxxxxxxxxxx>
Link: https://lore.kernel.org/r/20250611110800.458164-2-angelogioacchino.delregno@xxxxxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Chen Ni <nichen@xxxxxxxxxxx>
Date: Tue Jun 3 14:04:50 2025 +0800
@@ -52,9 +55,10 @@ Date: Tue Jun 3 14:04:50 2025 +0800
Compile tested only.
Signed-off-by: Chen Ni <nichen@xxxxxxxxxxx>
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Tested-by: Caleb James DeLisle <cjd@xxxxxxxx>
Link: https://lore.kernel.org/r/20250603060450.1310204-1-nichen@xxxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Frank Li <Frank.Li@xxxxxxx>
Date: Wed May 28 12:53:50 2025 -0400
@@ -67,9 +71,10 @@ Date: Wed May 28 12:53:50 2025 -0400
devices, which have existed for over 15 years.
Signed-off-by: Frank Li <Frank.Li@xxxxxxx>
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Acked-by: Conor Dooley <conor.dooley@xxxxxxxxxxxxx>
Link: https://lore.kernel.org/r/20250528165351.691848-1-Frank.Li@xxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Arnd Bergmann <arnd@xxxxxxxx>
Date: Fri Jun 20 13:19:35 2025 +0200
@@ -79,19 +84,21 @@ Date: Fri Jun 20 13:19:35 2025 +0200
The newly added function causes a build failure on 32-bit targets with
older compiler version such as gcc-10:
- arm-linux-gnueabi-ld: drivers/clocksource/timer-tegra186.o: in function `tegra186_wdt_get_timeleft':
- timer-tegra186.c:(.text+0x3c2): undefined reference to `__aeabi_uldivmod'
+ arm-linux-gnueabi-ld: drivers/clocksource/timer-tegra186.o: in function `tegra186_wdt_get_timeleft':
+ timer-tegra186.c:(.text+0x3c2): undefined reference to `__aeabi_uldivmod'
The calculation can trivially be changed to avoid the division entirely,
as USEC_PER_SEC is a multiple of 5. Change both such calculation for
consistency, even though gcc apparently managed to optimize the other one
properly already.
+ [ dlezcano : Fixed conflict with 20250614175556.922159-2-linux@xxxxxxxxxxxx ]
+
Fixes: 28c842c8b0f5 ("clocksource/drivers/timer-tegra186: Add WDIOC_GETTIMELEFT support")
Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
- Link: https://lore.kernel.org/r/20250620111939.3395525-1-arnd@xxxxxxxxxx
- [dlezcano] : Fixed conflict with 20250614175556.922159-2-linux@xxxxxxxxxxxx
Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
+ Link: https://lore.kernel.org/r/20250620111939.3395525-1-arnd@xxxxxxxxxx
Author: Guenter Roeck <linux@xxxxxxxxxxxx>
Date: Sat Jun 14 10:55:56 2025 -0700
@@ -102,35 +109,37 @@ Date: Sat Jun 14 10:55:56 2025 -0700
remaining watchdog timeout. Simplify to use 32-bit operations,
and add comments explaining why there will be no overflow.
- Cc: Pohsun Su <pohsuns@xxxxxxxxxx>
- Cc: Robert Lin <robelin@xxxxxxxxxx>
Signed-off-by: Guenter Roeck <linux@xxxxxxxxxxxx>
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Reviewed-by: Jon Hunter <jonathanh@xxxxxxxxxx>
+ Cc: Pohsun Su <pohsuns@xxxxxxxxxx>
+ Cc: Robert Lin <robelin@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20250614175556.922159-2-linux@xxxxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Guenter Roeck <linux@xxxxxxxxxxxx>
Date: Sat Jun 14 10:55:55 2025 -0700
clocksource/drivers/timer-tegra186: Avoid 64-bit divide operation
- Building the driver on xtensa fails with
+ Building the driver on xtensa fails with:
- tensa-linux-ld: drivers/clocksource/timer-tegra186.o:
+ tensa-linux-ld: drivers/clocksource/timer-tegra186.o:
in function `tegra186_timer_remove':
- timer-tegra186.c:(.text+0x350):
+ timer-tegra186.c:(.text+0x350):
undefined reference to `__udivdi3'
Avoid the problem by rearranging the offending code to avoid the 64-bit
divide operation.
Fixes: 28c842c8b0f5 ("clocksource/drivers/timer-tegra186: Add WDIOC_GETTIMELEFT support")
- Cc: Pohsun Su <pohsuns@xxxxxxxxxx>
- Cc: Robert Lin <robelin@xxxxxxxxxx>
Signed-off-by: Guenter Roeck <linux@xxxxxxxxxxxx>
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Reviewed-by: Jon Hunter <jonathanh@xxxxxxxxxx>
+ Cc: Pohsun Su <pohsuns@xxxxxxxxxx>
+ Cc: Robert Lin <robelin@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20250614175556.922159-1-linux@xxxxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Will McVicker <willmcvicker@xxxxxxxxxx>
Date: Fri Jun 20 11:17:09 2025 -0700
@@ -143,11 +152,12 @@ Date: Fri Jun 20 11:17:09 2025 -0700
automatically. This allows platforms like Android to build the driver as
a module if desired.
- Reviewed-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
- Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
- Link: https://lore.kernel.org/r/20250620181719.1399856-7-willmcvicker@xxxxxxxxxx
Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
+ Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
+ Reviewed-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
+ Link: https://lore.kernel.org/r/20250620181719.1399856-7-willmcvicker@xxxxxxxxxx
Author: Donghoon Yu <hoony.yu@xxxxxxxxxxx>
Date: Fri Jun 20 11:17:08 2025 -0700
@@ -159,19 +169,21 @@ Date: Fri Jun 20 11:17:08 2025 -0700
tick timer. Once the MCT driver is loaded, it can be used as the wakeup
source for the arch_timer.
+ Original commit from:
+
+ https://android.googlesource.com/kernel/gs/+/8a52a8288ec7d88ff78f0b37480dbb0e9c65bbfd]
+
Signed-off-by: Donghoon Yu <hoony.yu@xxxxxxxxxxx>
Signed-off-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
- [original commit from https://android.googlesource.com/kernel/gs/+/8a52a8288ec7d88ff78f0b37480dbb0e9c65bbfd]
- Reviewed-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
- Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
- Link: https://lore.kernel.org/r/20250620181719.1399856-6-willmcvicker@xxxxxxxxxx
Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
+ Link: https://lore.kernel.org/r/20250620181719.1399856-6-willmcvicker@xxxxxxxxxx
Author: Will McVicker <willmcvicker@xxxxxxxxxx>
Date: Fri Jun 20 11:17:07 2025 -0700
- clocksource/drivers/exynos_mct: Fix uninitialized irq name warning
+ clocksource/drivers/exynos_mct: Fix uninitialized IRQ name warning
The Exynos MCT driver doesn't set the clocksource name until the CPU
hotplug state is setup which happens after the IRQs are requested. This
@@ -217,32 +229,37 @@ Date: Fri Jun 20 11:17:07 2025 -0700
[ T430] load_module+0x1de0/0x2500
[ T430] init_module_from_file+0x8c/0xdc
+ Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
+ Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
Reviewed-by: Peter Griffin <peter.griffin@xxxxxxxxxx>
Reviewed-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
- Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
- Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20250620181719.1399856-5-willmcvicker@xxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Hosung Kim <hosung0.kim@xxxxxxxxxxx>
Date: Fri Jun 20 11:17:06 2025 -0700
clocksource/drivers/exynos_mct: Set local timer interrupts as percpu
- To allow the CPU to handle it's own clock events, we need to set the
+ To allow the CPU to handle its own clock events, we need to set the
IRQF_PERCPU flag. This prevents the local timer interrupts from
migrating to other CPUs.
+ Original commit from:
+
+ https://android.googlesource.com/kernel/gs/+/03267fad19f093bac979ca78309483e9eb3a8d16
+
Signed-off-by: Hosung Kim <hosung0.kim@xxxxxxxxxxx>
- [Original commit from https://android.googlesource.com/kernel/gs/+/03267fad19f093bac979ca78309483e9eb3a8d16]
+ Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
+ Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
Reviewed-by: Peter Griffin <peter.griffin@xxxxxxxxxx>
Reviewed-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
- Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
- Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20250620181719.1399856-4-willmcvicker@xxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
-Author: Will McVicker <willmcvicker@xxxxxxxxxx>
+Author: Donghoon Yu <hoony.yu@xxxxxxxxxxx>
Date: Fri Jun 20 11:17:05 2025 -0700
clocksource/drivers/exynos_mct: Don't register as a sched_clock on arm64
@@ -257,42 +274,44 @@ Date: Fri Jun 20 11:17:05 2025 -0700
Since ARM32 SoCs don't have an architected timer, the MCT must continue
to be used for sched_clock. Detailed discussion on this topic can be
- found at [1].
+ found at:
+
+ https://lore.kernel.org/linux-samsung-soc/1400188079-21832-1-git-send-email-chirantan@xxxxxxxxxxxx/
- [1] https://lore.kernel.org/linux-samsung-soc/1400188079-21832-1-git-send-email-chirantan@xxxxxxxxxxxx/
+ Original commit from:
- [Original commit from https://android.googlesource.com/kernel/gs/+/630817f7080e92c5e0216095ff52f6eb8dd00727
+ https://android.googlesource.com/kernel/gs/+/630817f7080e92c5e0216095ff52f6eb8dd00727
Signed-off-by: Donghoon Yu <hoony.yu@xxxxxxxxxxx>
Signed-off-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
- Reviewed-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
- Acked-by: John Stultz <jstultz@xxxxxxxxxx>
- Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
- Link: https://lore.kernel.org/r/20250620181719.1399856-3-willmcvicker@xxxxxxxxxx
Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
+ Acked-by: John Stultz <jstultz@xxxxxxxxxx>
+ Link: https://lore.kernel.org/r/20250620181719.1399856-3-willmcvicker@xxxxxxxxxx
Author: Will McVicker <willmcvicker@xxxxxxxxxx>
Date: Fri Jun 20 11:17:04 2025 -0700
- of/irq: Export of_irq_count for modules
+ of/irq: Export of_irq_count() for modules
- Need to export `of_irq_count` in preparation for modularizing the Exynos
+ Need to export of_irq_count() in preparation for modularizing the Exynos
MCT driver which uses this API for setting up the timer IRQs.
- Acked-by: Rob Herring (Arm) <robh@xxxxxxxxxx>
- Acked-by: Arnd Bergmann <arnd@xxxxxxxx>
+ Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
+ Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
Reviewed-by: Linus Walleij <linus.walleij-QSEj5FYQhm4dnm+yROfE0A@xxxxxxxxxxxxxxxx>
Reviewed-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
- Tested-by: Youngmin Nam <youngmin.nam@xxxxxxxxxxx>
- Signed-off-by: Will McVicker <willmcvicker@xxxxxxxxxx>
+ Acked-by: Rob Herring (Arm) <robh@xxxxxxxxxx>
+ Acked-by: Arnd Bergmann <arnd@xxxxxxxx>
Link: https://lore.kernel.org/r/20250620181719.1399856-2-willmcvicker@xxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Ben Zong-You Xie <ben717@xxxxxxxxxxxxx>
Date: Fri Jul 11 21:30:21 2025 +0800
- dt-bindings: timer: add Andes machine timer
+ dt-bindings: timer: Add Andes machine timer
Add the DT binding documentation for Andes machine timer.
@@ -301,43 +320,47 @@ Date: Fri Jul 11 21:30:21 2025 +0800
the implementation of the machine timer, and it contains memory-mapped
registers (mtime and mtimecmp). This device supports up to 32 cores.
- Acked-by: Conor Dooley <conor.dooley@xxxxxxxxxxxxx>
Signed-off-by: Ben Zong-You Xie <ben717@xxxxxxxxxxxxx>
- Link: https://lore.kernel.org/r/20250711133025.2192404-6-ben717@xxxxxxxxxxxxx
Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
+ Acked-by: Conor Dooley <conor.dooley@xxxxxxxxxxxxx>
+ Link: https://lore.kernel.org/r/20250711133025.2192404-6-ben717@xxxxxxxxxxxxx
Author: Frank Li <Frank.Li@xxxxxxx>
Date: Fri May 23 10:14:37 2025 -0400
- dt-bindings: timer: fsl,ftm-timer: use items for reg
+ dt-bindings: timer: fsl,ftm-timer: Use 'items' for 'reg'
The original txt binding doc is:
+
reg : Specifies base physical address and size of the register sets for
the clock event device and clock source device.
- And existed dts provide two reg MMIO spaces. So change to use items to
- descript reg property.
+ And existing DTS drivers provide two 'reg' MMIO spaces. So change
+ this driver to use 'items' to describe the 'reg' property.
- Update examples.
+ Update examples as well.
Fixes: 8fc30d8f8e86 ("dt-bindings: timer: fsl,ftm-timer: Convert to dtschema")
- Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
Signed-off-by: Frank Li <Frank.Li@xxxxxxx>
- Link: https://lore.kernel.org/r/20250523141437.533643-1-Frank.Li@xxxxxxx
Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
+ Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
+ Link: https://lore.kernel.org/r/20250523141437.533643-1-Frank.Li@xxxxxxx
Author: Max Shevchenko <wctrl@xxxxxxxxx>
Date: Wed Jul 2 13:50:40 2025 +0300
- dt-bindings: timer: mediatek: add MT6572
+ dt-bindings: timer: mediatek: Add MT6572
Add a compatible string for timer on the MT6572 SoC.
+ Signed-off-by: Max Shevchenko <wctrl@xxxxxxxxx>
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@xxxxxxxxxxxxx>
Acked-by: Rob Herring (Arm) <robh@xxxxxxxxxx>
- Signed-off-by: Max Shevchenko <wctrl@xxxxxxxxx>
Link: https://lore.kernel.org/r/20250702-mt6572-v4-3-bde75b7ed445@xxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Rob Herring (Arm) <robh@xxxxxxxxxx>
Date: Wed Jun 11 18:26:20 2025 -0500
@@ -349,30 +372,32 @@ Date: Wed Jun 11 18:26:20 2025 -0500
interrupts can also be anywhere from 1 to 8. The clock-names order was
reversed compared to what's used.
- Reviewed-by: Linus Walleij <linus.walleij@xxxxxxxxxx>
Signed-off-by: Rob Herring (Arm) <robh@xxxxxxxxxx>
- Link: https://lore.kernel.org/r/20250611232621.1508116-1-robh@xxxxxxxxxx
Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
+ Reviewed-by: Linus Walleij <linus.walleij@xxxxxxxxxx>
+ Link: https://lore.kernel.org/r/20250611232621.1508116-1-robh@xxxxxxxxxx
Author: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Date: Mon Jun 2 17:18:51 2025 +0200
- time/sched_clock: Export symbol for sched_clock register function
+ time/sched_clock: Export symbol for sched_clock_register() function
- The timer drivers could be converted into modules. The different
+ Timer drivers could be converted into modules. The different
functions to register the clocksource or the clockevent are already
- exporting their symbols for modules but the sched_clock_register()
+ exporting their symbols for modules, but the sched_clock_register()
function is missing.
- Export the symbols so the drivers using this function can be converted
+ Export the symbol so the drivers using this function can be converted
into modules.
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Reviewed-by: Will McVicker <willmcvicker@xxxxxxxxxx>
- Acked-by: John Stultz <jstultz@xxxxxxxxxx>
Reviewed-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Reviewed-by: Carlos Llamas <cmllamas@xxxxxxxxxx>
+ Acked-by: John Stultz <jstultz@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20250602151853.1942521-8-daniel.lezcano@xxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Date: Mon Jun 2 17:18:50 2025 +0200
@@ -381,7 +406,7 @@ Date: Mon Jun 2 17:18:50 2025 +0200
The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
- especially true with the clockevents where there is no function to
+ especially true with clockevents where there is no function to
unregister them.
The core time framework correctly handles the module refcount with the
@@ -391,9 +416,10 @@ Date: Mon Jun 2 17:18:50 2025 +0200
stupid things happening when the driver will be converted into a
module.
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Reviewed-by: Will McVicker <willmcvicker@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20250602151853.1942521-7-daniel.lezcano@xxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Date: Mon Jun 2 17:18:49 2025 +0200
@@ -402,7 +428,7 @@ Date: Mon Jun 2 17:18:49 2025 +0200
The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
- especially true with the clockevents where there is no function to
+ especially true with clockevents where there is no function to
unregister them.
The core time framework correctly handles the module refcount with the
@@ -412,9 +438,10 @@ Date: Mon Jun 2 17:18:49 2025 +0200
stupid things happening when the driver will be converted into a
module.
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Reviewed-by: Will McVicker <willmcvicker@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20250602151853.1942521-6-daniel.lezcano@xxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Date: Mon Jun 2 17:18:48 2025 +0200
@@ -423,7 +450,7 @@ Date: Mon Jun 2 17:18:48 2025 +0200
The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
- especially true with the clockevents where there is no function to
+ especially true with clockevents where there is no function to
unregister them.
The core time framework correctly handles the module refcount with the
@@ -433,9 +460,10 @@ Date: Mon Jun 2 17:18:48 2025 +0200
stupid things happening when the driver will be converted into a
module.
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Reviewed-by: Will McVicker <willmcvicker@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20250602151853.1942521-5-daniel.lezcano@xxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Date: Mon Jun 2 17:18:47 2025 +0200
@@ -444,7 +472,7 @@ Date: Mon Jun 2 17:18:47 2025 +0200
The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
- especially true with the clockevents where there is no function to
+ especially true with clockevents where there is no function to
unregister them.
The core time framework correctly handles the module refcount with the
@@ -454,10 +482,11 @@ Date: Mon Jun 2 17:18:47 2025 +0200
stupid things happening when the driver will be converted into a
module.
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Reviewed-by: Will McVicker <willmcvicker@xxxxxxxxxx>
Acked-by: Chen-Yu Tsai <wens@xxxxxxxx>
Link: https://lore.kernel.org/r/20250602151853.1942521-4-daniel.lezcano@xxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Date: Mon Jun 2 17:18:46 2025 +0200
@@ -466,7 +495,7 @@ Date: Mon Jun 2 17:18:46 2025 +0200
The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
- especially true with the clockevents where there is no function to
+ especially true with clockevents where there is no function to
unregister them.
The core time framework correctly handles the module refcount with the
@@ -476,9 +505,10 @@ Date: Mon Jun 2 17:18:46 2025 +0200
stupid things happening when the driver will be converted into a
module.
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Reviewed-by: Will McVicker <willmcvicker@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20250602151853.1942521-3-daniel.lezcano@xxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Author: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
Date: Mon Jun 2 17:18:45 2025 +0200
@@ -487,7 +517,7 @@ Date: Mon Jun 2 17:18:45 2025 +0200
The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
- especially true with the clockevents where there is no function to
+ especially true with clockevents where there is no function to
unregister them.
The core time framework correctly handles the module refcount with the
@@ -497,6 +527,7 @@ Date: Mon Jun 2 17:18:45 2025 +0200
stupid things happening when the driver will be converted into a
module.
+ Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>
+ Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
Reviewed-by: Will McVicker <willmcvicker@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20250602151853.1942521-2-daniel.lezcano@xxxxxxxxxx
- Signed-off-by: Daniel Lezcano <daniel.lezcano@xxxxxxxxxx>