On 03/17/19 13:41, Oskari Lemmelä wrote:
On 3/17/19 6:04 PM, Peter Robinson wrote:
On Sun, Mar 17, 2019 at 2:56 PM Oskari Lemmela oskari@lemmela.net wrote:
Fixes spurious timeouts which have been seen during testing SPI_SUNXI driver. The false timeouts disappear when number of bits reduced to 10 in workaround.
The false timeouts are caused by timer backward jumps.
Wouldn't it be best to apply the same fix as the kernel uses here [1] as a more permanent fix?
This _is_ the same fix as the kernel uses.
Sure. Difference is that retry counter was added to while loop in kernel side.
And the retry counter isn't needed here, since there is no preemption in u-boot.
Only CNTPCT is used in u-boot. So CNTVCT part of kernel patch is currently not needed?
Correct.
Signed-off-by: Oskari Lemmela oskari@lemmela.net
arch/arm/cpu/armv8/generic_timer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/arm/cpu/armv8/generic_timer.c b/arch/arm/cpu/armv8/generic_timer.c index c1706dcec1..2e06ee4ed2 100644 --- a/arch/arm/cpu/armv8/generic_timer.c +++ b/arch/arm/cpu/armv8/generic_timer.c @@ -66,7 +66,7 @@ unsigned long timer_read_counter(void) isb(); do { asm volatile("mrs %0, cntpct_el0" : "=r" (cntpct)); - } while (((cntpct + 1) & GENMASK(10, 0)) <= 1); + } while (((cntpct + 1) & GENMASK(9, 0)) <= 1);
Please update the 11 to 10 in the comment above as well.
return cntpct; } -- 2.17.1
Thanks, Samuel