[Qemu-devel] [PATCH 1/2] arm_timer: reload timer when enabled

[Qemu-devel] [PATCH 1/2] arm_timer: reload timer when enabled

[Rabin Vincent]

Reload the timer when TimerControl is written, if the timer is to be
enabled.  Otherwise, if an earlier write to TimerLoad was done while
periodic mode was not set, s->delta may incorrectly still have the value
of the maximum limit instead of the value written to TimerLoad.

This problem is evident on versatileap on current linux-next, which
enables TIMER_CTRL_32BIT before writing to TimerLoad and then enabling
periodic mode and starting the timer.  This causes the first periodic
tick to be scheduled to occur after 0xffffffff periods, leading to a
perceived hang while the kernel waits for the first timer tick.

Signed-off-by: Rabin Vincent <rabin@rab.in>
---
 hw/arm_timer.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/hw/arm_timer.c b/hw/arm_timer.c
index 9fef191..5b6947a 100644
--- a/hw/arm_timer.c
+++ b/hw/arm_timer.c
@@ -113,7 +113,7 @@ static void arm_timer_write(void *opaque, target_phys_addr_t offset,
         case 1: freq >>= 4; break;
         case 2: freq >>= 8; break;
         }
-        arm_timer_recalibrate(s, 0);
+        arm_timer_recalibrate(s, s->control & TIMER_CTRL_ENABLE);
         ptimer_set_freq(s->timer, freq);
         if (s->control & TIMER_CTRL_ENABLE) {
             /* Restart the timer if still enabled.  */
-- 
1.7.0.4

[Qemu-devel] [PATCH 2/2] arm_timer: fix oneshot mode

[Rabin Vincent]

In oneshot mode, the delta needs to come from the TimerLoad register,
not the maximum limit.

Signed-off-by: Rabin Vincent <rabin@rab.in>
---
 hw/arm_timer.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/hw/arm_timer.c b/hw/arm_timer.c
index 5b6947a..9073ffc 100644
--- a/hw/arm_timer.c
+++ b/hw/arm_timer.c
@@ -71,7 +71,7 @@ static void arm_timer_recalibrate(arm_timer_state *s, int reload)
 {
     uint32_t limit;
 
-    if ((s->control & TIMER_CTRL_PERIODIC) == 0) {
+    if ((s->control & (TIMER_CTRL_PERIODIC | TIMER_CTRL_ONESHOT)) == 0) {
         /* Free running.  */
         if (s->control & TIMER_CTRL_32BIT)
             limit = 0xffffffff;
-- 
1.7.0.4

Re: [Qemu-devel] [PATCH 2/2] arm_timer: fix oneshot mode

[Aurelien Jarno]

On Sun, May 02, 2010 at 03:20:52PM +0530, Rabin Vincent wrote:
> In oneshot mode, the delta needs to come from the TimerLoad register,
> not the maximum limit.
> 
> Signed-off-by: Rabin Vincent <rabin@rab.in>

Thanks, applied.

> ---
>  hw/arm_timer.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/hw/arm_timer.c b/hw/arm_timer.c
> index 5b6947a..9073ffc 100644
> --- a/hw/arm_timer.c
> +++ b/hw/arm_timer.c
> @@ -71,7 +71,7 @@ static void arm_timer_recalibrate(arm_timer_state *s, int reload)
>  {
>      uint32_t limit;
>  
> -    if ((s->control & TIMER_CTRL_PERIODIC) == 0) {
> +    if ((s->control & (TIMER_CTRL_PERIODIC | TIMER_CTRL_ONESHOT)) == 0) {
>          /* Free running.  */
>          if (s->control & TIMER_CTRL_32BIT)
>              limit = 0xffffffff;
> -- 
> 1.7.0.4
> 
> 
> 
> 

-- 
Aurelien Jarno                          GPG: 1024D/F1BCDB73
aurelien@aurel32.net                 http://www.aurel32.net

[Qemu-devel] Re: [PATCH 1/2] arm_timer: reload timer when enabled

[Rabin Vincent]

On Sun, May 02, 2010 at 03:20:51PM +0530, Rabin Vincent wrote:
> Reload the timer when TimerControl is written, if the timer is to be
> enabled.  Otherwise, if an earlier write to TimerLoad was done while
> periodic mode was not set, s->delta may incorrectly still have the value
> of the maximum limit instead of the value written to TimerLoad.
> 
> This problem is evident on versatileap on current linux-next, which
> enables TIMER_CTRL_32BIT before writing to TimerLoad and then enabling
> periodic mode and starting the timer.  This causes the first periodic
> tick to be scheduled to occur after 0xffffffff periods, leading to a
> perceived hang while the kernel waits for the first timer tick.
> 
> Signed-off-by: Rabin Vincent <rabin@rab.in>

Could these patches please be applied?  What was then linux-next is now
current Linux mainline, and it doesn't boot without this patch.

Rabin

Re: [Qemu-devel] [PATCH 1/2] arm_timer: reload timer when enabled

[Aurelien Jarno]

On Sun, May 02, 2010 at 03:20:51PM +0530, Rabin Vincent wrote:
> Reload the timer when TimerControl is written, if the timer is to be
> enabled.  Otherwise, if an earlier write to TimerLoad was done while
> periodic mode was not set, s->delta may incorrectly still have the value
> of the maximum limit instead of the value written to TimerLoad.
> 
> This problem is evident on versatileap on current linux-next, which
> enables TIMER_CTRL_32BIT before writing to TimerLoad and then enabling
> periodic mode and starting the timer.  This causes the first periodic
> tick to be scheduled to occur after 0xffffffff periods, leading to a
> perceived hang while the kernel waits for the first timer tick.
> 
> Signed-off-by: Rabin Vincent <rabin@rab.in>

Thanks, applied.

> ---
>  hw/arm_timer.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/hw/arm_timer.c b/hw/arm_timer.c
> index 9fef191..5b6947a 100644
> --- a/hw/arm_timer.c
> +++ b/hw/arm_timer.c
> @@ -113,7 +113,7 @@ static void arm_timer_write(void *opaque, target_phys_addr_t offset,
>          case 1: freq >>= 4; break;
>          case 2: freq >>= 8; break;
>          }
> -        arm_timer_recalibrate(s, 0);
> +        arm_timer_recalibrate(s, s->control & TIMER_CTRL_ENABLE);
>          ptimer_set_freq(s->timer, freq);
>          if (s->control & TIMER_CTRL_ENABLE) {
>              /* Restart the timer if still enabled.  */
> -- 
> 1.7.0.4
> 
> 
> 
> 

-- 
Aurelien Jarno                          GPG: 1024D/F1BCDB73
aurelien@aurel32.net                 http://www.aurel32.net