* [PATCH 1/3] timer-uclass: add timer_get_ops() macro
@ 2022-10-12 5:36 Ovidiu Panait
2022-10-12 5:36 ` [PATCH 2/3] timer-uclass: relocate ops pointers for CONFIG_NEEDS_MANUAL_RELOC Ovidiu Panait
` (2 more replies)
0 siblings, 3 replies; 8+ messages in thread
From: Ovidiu Panait @ 2022-10-12 5:36 UTC (permalink / raw)
To: u-boot; +Cc: Ovidiu Panait, Andy Shevchenko, Patrice Chotard, Simon Glass
Align timer uclass with the other subsystems and provide a timer_get_ops()
convenience macro.
Using this instead of the generic device_get_ops() also prevents
-Wdiscarded-qualifiers warnings when used with non-const variables.
Signed-off-by: Ovidiu Panait <ovpanait@gmail.com>
---
drivers/timer/timer-uclass.c | 2 +-
include/timer.h | 2 ++
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/timer/timer-uclass.c b/drivers/timer/timer-uclass.c
index cbc3647698..bdc77b3822 100644
--- a/drivers/timer/timer-uclass.c
+++ b/drivers/timer/timer-uclass.c
@@ -32,7 +32,7 @@ DECLARE_GLOBAL_DATA_PTR;
int notrace timer_get_count(struct udevice *dev, u64 *count)
{
- const struct timer_ops *ops = device_get_ops(dev);
+ struct timer_ops *ops = timer_get_ops(dev);
if (!ops->get_count)
return -ENOSYS;
diff --git a/include/timer.h b/include/timer.h
index a044cb034e..d33a26e28f 100644
--- a/include/timer.h
+++ b/include/timer.h
@@ -6,6 +6,8 @@
#ifndef _TIMER_H_
#define _TIMER_H_
+#define timer_get_ops(dev) ((struct timer_ops *)(dev)->driver->ops)
+
/**
* dm_timer_init() - initialize a timer for time keeping. On success
* initializes gd->timer so that lib/timer can use it for future
--
2.25.1
^ permalink raw reply related [flat|nested] 8+ messages in thread* [PATCH 2/3] timer-uclass: relocate ops pointers for CONFIG_NEEDS_MANUAL_RELOC
2022-10-12 5:36 [PATCH 1/3] timer-uclass: add timer_get_ops() macro Ovidiu Panait
@ 2022-10-12 5:36 ` Ovidiu Panait
2022-10-12 12:59 ` Simon Glass
2022-10-12 5:36 ` [PATCH 3/3] timer: xilinx-timer: use timer_conv_64() to fix timer wrap around Ovidiu Panait
2022-10-12 12:59 ` [PATCH 1/3] timer-uclass: add timer_get_ops() macro Simon Glass
2 siblings, 1 reply; 8+ messages in thread
From: Ovidiu Panait @ 2022-10-12 5:36 UTC (permalink / raw)
To: u-boot; +Cc: Ovidiu Panait, Andy Shevchenko, Patrice Chotard, Simon Glass
Relocate timer_ops pointers when CONFIG_NEEDS_MANUAL_RELOC is enabled.
The (gd->flags & GD_FLG_RELOC) check was added to make sure the reloc_done
logic works for drivers that use DM_FLAG_PRE_RELOC.
Signed-off-by: Ovidiu Panait <ovpanait@gmail.com>
---
drivers/timer/timer-uclass.c | 14 ++++++++++++++
1 file changed, 14 insertions(+)
diff --git a/drivers/timer/timer-uclass.c b/drivers/timer/timer-uclass.c
index bdc77b3822..bb71979213 100644
--- a/drivers/timer/timer-uclass.c
+++ b/drivers/timer/timer-uclass.c
@@ -18,6 +18,7 @@
#include <init.h>
#include <timer.h>
#include <linux/err.h>
+#include <relocate.h>
DECLARE_GLOBAL_DATA_PTR;
@@ -50,6 +51,19 @@ unsigned long notrace timer_get_rate(struct udevice *dev)
static int timer_pre_probe(struct udevice *dev)
{
+ if (IS_ENABLED(CONFIG_NEEDS_MANUAL_RELOC) &&
+ (gd->flags & GD_FLG_RELOC)) {
+ struct timer_ops *ops = timer_get_ops(dev);
+ static int reloc_done;
+
+ if (!reloc_done) {
+ if (ops->get_count)
+ MANUAL_RELOC(ops->get_count);
+
+ reloc_done++;
+ }
+ }
+
if (CONFIG_IS_ENABLED(OF_REAL)) {
struct timer_dev_priv *uc_priv = dev_get_uclass_priv(dev);
struct clk timer_clk;
--
2.25.1
^ permalink raw reply related [flat|nested] 8+ messages in thread* Re: [PATCH 2/3] timer-uclass: relocate ops pointers for CONFIG_NEEDS_MANUAL_RELOC
2022-10-12 5:36 ` [PATCH 2/3] timer-uclass: relocate ops pointers for CONFIG_NEEDS_MANUAL_RELOC Ovidiu Panait
@ 2022-10-12 12:59 ` Simon Glass
2022-10-12 15:21 ` Ovidiu Panait
0 siblings, 1 reply; 8+ messages in thread
From: Simon Glass @ 2022-10-12 12:59 UTC (permalink / raw)
To: Ovidiu Panait; +Cc: u-boot, Andy Shevchenko, Patrice Chotard
On Tue, 11 Oct 2022 at 23:37, Ovidiu Panait <ovpanait@gmail.com> wrote:
>
> Relocate timer_ops pointers when CONFIG_NEEDS_MANUAL_RELOC is enabled.
>
> The (gd->flags & GD_FLG_RELOC) check was added to make sure the reloc_done
> logic works for drivers that use DM_FLAG_PRE_RELOC.
>
> Signed-off-by: Ovidiu Panait <ovpanait@gmail.com>
> ---
>
> drivers/timer/timer-uclass.c | 14 ++++++++++++++
> 1 file changed, 14 insertions(+)
Reviewed-by: Simon Glass <sjg@chromium.org>
Which (broken) toolchain are you using?
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH 2/3] timer-uclass: relocate ops pointers for CONFIG_NEEDS_MANUAL_RELOC
2022-10-12 12:59 ` Simon Glass
@ 2022-10-12 15:21 ` Ovidiu Panait
0 siblings, 0 replies; 8+ messages in thread
From: Ovidiu Panait @ 2022-10-12 15:21 UTC (permalink / raw)
To: Simon Glass; +Cc: u-boot, Andy Shevchenko, Patrice Chotard
Hi Simon,
On 10/12/22 15:59, Simon Glass wrote:
> On Tue, 11 Oct 2022 at 23:37, Ovidiu Panait <ovpanait@gmail.com> wrote:
>> Relocate timer_ops pointers when CONFIG_NEEDS_MANUAL_RELOC is enabled.
>>
>> The (gd->flags & GD_FLG_RELOC) check was added to make sure the reloc_done
>> logic works for drivers that use DM_FLAG_PRE_RELOC.
>>
>> Signed-off-by: Ovidiu Panait <ovpanait@gmail.com>
>> ---
>>
>> drivers/timer/timer-uclass.c | 14 ++++++++++++++
>> 1 file changed, 14 insertions(+)
> Reviewed-by: Simon Glass <sjg@chromium.org>
>
> Which (broken) toolchain are you using?
Microblaze and m68k still use CONFIG_NEEDS_MANUAL_RELOC=y by default. I
was testing on microblaze.
Ovidiu
^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH 3/3] timer: xilinx-timer: use timer_conv_64() to fix timer wrap around
2022-10-12 5:36 [PATCH 1/3] timer-uclass: add timer_get_ops() macro Ovidiu Panait
2022-10-12 5:36 ` [PATCH 2/3] timer-uclass: relocate ops pointers for CONFIG_NEEDS_MANUAL_RELOC Ovidiu Panait
@ 2022-10-12 5:36 ` Ovidiu Panait
2022-10-12 8:58 ` Michal Simek
2022-10-25 9:32 ` Michal Simek
2022-10-12 12:59 ` [PATCH 1/3] timer-uclass: add timer_get_ops() macro Simon Glass
2 siblings, 2 replies; 8+ messages in thread
From: Ovidiu Panait @ 2022-10-12 5:36 UTC (permalink / raw)
To: u-boot; +Cc: Ovidiu Panait, Michal Simek
Current xilinx_timer_get_count() implementation does not take into account
the periodic 32-bit wrap arounds, as it directly returns the 32-bit counter
register value. The roll-overs cause problems in the upper timer layers, as
generic timer code expects an incrementing 64-bit value from get_count() to
work correctly.
Add the missing 64-bit up-conversion to fix random hangs/delays in
__udelay().
Fixes: a36d86720f ("microblaze: Convert axi timer to DM driver")
Signed-off-by: Ovidiu Panait <ovpanait@gmail.com>
---
drivers/timer/xilinx-timer.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/timer/xilinx-timer.c b/drivers/timer/xilinx-timer.c
index 75b4473b63..172fd9f929 100644
--- a/drivers/timer/xilinx-timer.c
+++ b/drivers/timer/xilinx-timer.c
@@ -40,7 +40,7 @@ static u64 xilinx_timer_get_count(struct udevice *dev)
regmap_read(priv->regs, TIMER_COUNTER_OFFSET, &value);
- return value;
+ return timer_conv_64(value);
}
static int xilinx_timer_probe(struct udevice *dev)
--
2.25.1
^ permalink raw reply related [flat|nested] 8+ messages in thread* Re: [PATCH 3/3] timer: xilinx-timer: use timer_conv_64() to fix timer wrap around
2022-10-12 5:36 ` [PATCH 3/3] timer: xilinx-timer: use timer_conv_64() to fix timer wrap around Ovidiu Panait
@ 2022-10-12 8:58 ` Michal Simek
2022-10-25 9:32 ` Michal Simek
1 sibling, 0 replies; 8+ messages in thread
From: Michal Simek @ 2022-10-12 8:58 UTC (permalink / raw)
To: Ovidiu Panait, u-boot
On 10/12/22 07:36, Ovidiu Panait wrote:
> Current xilinx_timer_get_count() implementation does not take into account
> the periodic 32-bit wrap arounds, as it directly returns the 32-bit counter
> register value. The roll-overs cause problems in the upper timer layers, as
> generic timer code expects an incrementing 64-bit value from get_count() to
> work correctly.
>
> Add the missing 64-bit up-conversion to fix random hangs/delays in
> __udelay().
>
> Fixes: a36d86720f ("microblaze: Convert axi timer to DM driver")
> Signed-off-by: Ovidiu Panait <ovpanait@gmail.com>
> ---
>
> drivers/timer/xilinx-timer.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/timer/xilinx-timer.c b/drivers/timer/xilinx-timer.c
> index 75b4473b63..172fd9f929 100644
> --- a/drivers/timer/xilinx-timer.c
> +++ b/drivers/timer/xilinx-timer.c
> @@ -40,7 +40,7 @@ static u64 xilinx_timer_get_count(struct udevice *dev)
>
> regmap_read(priv->regs, TIMER_COUNTER_OFFSET, &value);
>
> - return value;
> + return timer_conv_64(value);
> }
>
> static int xilinx_timer_probe(struct udevice *dev)
Good catch.
Reviewed-by: Michal Simek <michal.simek@amd.com>
I will let Simon to comment other two patches.
Thanks,
Michal
^ permalink raw reply [flat|nested] 8+ messages in thread* Re: [PATCH 3/3] timer: xilinx-timer: use timer_conv_64() to fix timer wrap around
2022-10-12 5:36 ` [PATCH 3/3] timer: xilinx-timer: use timer_conv_64() to fix timer wrap around Ovidiu Panait
2022-10-12 8:58 ` Michal Simek
@ 2022-10-25 9:32 ` Michal Simek
1 sibling, 0 replies; 8+ messages in thread
From: Michal Simek @ 2022-10-25 9:32 UTC (permalink / raw)
To: Ovidiu Panait, u-boot
On 10/12/22 07:36, Ovidiu Panait wrote:
> Current xilinx_timer_get_count() implementation does not take into account
> the periodic 32-bit wrap arounds, as it directly returns the 32-bit counter
> register value. The roll-overs cause problems in the upper timer layers, as
> generic timer code expects an incrementing 64-bit value from get_count() to
> work correctly.
>
> Add the missing 64-bit up-conversion to fix random hangs/delays in
> __udelay().
>
> Fixes: a36d86720f ("microblaze: Convert axi timer to DM driver")
> Signed-off-by: Ovidiu Panait <ovpanait@gmail.com>
> ---
>
> drivers/timer/xilinx-timer.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/timer/xilinx-timer.c b/drivers/timer/xilinx-timer.c
> index 75b4473b63..172fd9f929 100644
> --- a/drivers/timer/xilinx-timer.c
> +++ b/drivers/timer/xilinx-timer.c
> @@ -40,7 +40,7 @@ static u64 xilinx_timer_get_count(struct udevice *dev)
>
> regmap_read(priv->regs, TIMER_COUNTER_OFFSET, &value);
>
> - return value;
> + return timer_conv_64(value);
> }
>
> static int xilinx_timer_probe(struct udevice *dev)
Applied all.
M
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH 1/3] timer-uclass: add timer_get_ops() macro
2022-10-12 5:36 [PATCH 1/3] timer-uclass: add timer_get_ops() macro Ovidiu Panait
2022-10-12 5:36 ` [PATCH 2/3] timer-uclass: relocate ops pointers for CONFIG_NEEDS_MANUAL_RELOC Ovidiu Panait
2022-10-12 5:36 ` [PATCH 3/3] timer: xilinx-timer: use timer_conv_64() to fix timer wrap around Ovidiu Panait
@ 2022-10-12 12:59 ` Simon Glass
2 siblings, 0 replies; 8+ messages in thread
From: Simon Glass @ 2022-10-12 12:59 UTC (permalink / raw)
To: Ovidiu Panait; +Cc: u-boot, Andy Shevchenko, Patrice Chotard
On Tue, 11 Oct 2022 at 23:37, Ovidiu Panait <ovpanait@gmail.com> wrote:
>
> Align timer uclass with the other subsystems and provide a timer_get_ops()
> convenience macro.
>
> Using this instead of the generic device_get_ops() also prevents
> -Wdiscarded-qualifiers warnings when used with non-const variables.
>
> Signed-off-by: Ovidiu Panait <ovpanait@gmail.com>
> ---
>
> drivers/timer/timer-uclass.c | 2 +-
> include/timer.h | 2 ++
> 2 files changed, 3 insertions(+), 1 deletion(-)
Reviewed-by: Simon Glass <sjg@chromium.org>
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2022-10-25 9:33 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2022-10-12 5:36 [PATCH 1/3] timer-uclass: add timer_get_ops() macro Ovidiu Panait
2022-10-12 5:36 ` [PATCH 2/3] timer-uclass: relocate ops pointers for CONFIG_NEEDS_MANUAL_RELOC Ovidiu Panait
2022-10-12 12:59 ` Simon Glass
2022-10-12 15:21 ` Ovidiu Panait
2022-10-12 5:36 ` [PATCH 3/3] timer: xilinx-timer: use timer_conv_64() to fix timer wrap around Ovidiu Panait
2022-10-12 8:58 ` Michal Simek
2022-10-25 9:32 ` Michal Simek
2022-10-12 12:59 ` [PATCH 1/3] timer-uclass: add timer_get_ops() macro Simon Glass
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox