* [PATCH v2] i2c: nvidia-gpu: Handle timeout correctly in gpu_i2c_check_status()
@ 2020-03-24 13:07 Kai-Heng Feng
2020-03-24 13:41 ` Andy Shevchenko
0 siblings, 1 reply; 3+ messages in thread
From: Kai-Heng Feng @ 2020-03-24 13:07 UTC (permalink / raw)
To: ajayg
Cc: Kai-Heng Feng, Andy Shevchenko, Wolfram Sang,
open list:I2C CONTROLLER DRIVER FOR NVIDIA GPU, open list
Nvidia card may come with a "phantom" UCSI device, and its driver gets
stuck in probe routine, prevents any system PM operations like suspend.
When the target time equals to jiffies, it's not included by
time_is_before_jiffies(). So let's use a boolean to make sure the
operation is done or timeout.
Fixes: c71bcdcb42a7 ("i2c: add i2c bus driver for NVIDIA GPU")
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
---
v2:
- Use a boolean to make sure the operation is timeout or not.
drivers/i2c/busses/i2c-nvidia-gpu.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/drivers/i2c/busses/i2c-nvidia-gpu.c b/drivers/i2c/busses/i2c-nvidia-gpu.c
index 62e18b4db0ed..bdb48600af0e 100644
--- a/drivers/i2c/busses/i2c-nvidia-gpu.c
+++ b/drivers/i2c/busses/i2c-nvidia-gpu.c
@@ -77,18 +77,20 @@ static int gpu_i2c_check_status(struct gpu_i2c_dev *i2cd)
{
unsigned long target = jiffies + msecs_to_jiffies(1000);
u32 val;
+ bool done = false;
do {
val = readl(i2cd->regs + I2C_MST_CNTL);
- if (!(val & I2C_MST_CNTL_CYCLE_TRIGGER))
- break;
- if ((val & I2C_MST_CNTL_STATUS) !=
- I2C_MST_CNTL_STATUS_BUS_BUSY)
+ if (!(val & I2C_MST_CNTL_CYCLE_TRIGGER)
+ || (val & I2C_MST_CNTL_STATUS) !=
+ I2C_MST_CNTL_STATUS_BUS_BUSY) {
+ done = true;
break;
+ }
usleep_range(500, 600);
} while (time_is_after_jiffies(target));
- if (time_is_before_jiffies(target)) {
+ if (!done) {
dev_err(i2cd->dev, "i2c timeout error %x\n", val);
return -ETIMEDOUT;
}
--
2.17.1
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH v2] i2c: nvidia-gpu: Handle timeout correctly in gpu_i2c_check_status()
2020-03-24 13:07 [PATCH v2] i2c: nvidia-gpu: Handle timeout correctly in gpu_i2c_check_status() Kai-Heng Feng
@ 2020-03-24 13:41 ` Andy Shevchenko
2020-03-24 14:39 ` Kai-Heng Feng
0 siblings, 1 reply; 3+ messages in thread
From: Andy Shevchenko @ 2020-03-24 13:41 UTC (permalink / raw)
To: Kai-Heng Feng
Cc: ajayg, Wolfram Sang,
open list:I2C CONTROLLER DRIVER FOR NVIDIA GPU, open list
On Tue, Mar 24, 2020 at 09:07:12PM +0800, Kai-Heng Feng wrote:
> Nvidia card may come with a "phantom" UCSI device, and its driver gets
> stuck in probe routine, prevents any system PM operations like suspend.
>
> When the target time equals to jiffies, it's not included by
> time_is_before_jiffies(). So let's use a boolean to make sure the
> operation is done or timeout.
Thank you for an update, my comments below.
> unsigned long target = jiffies + msecs_to_jiffies(1000);
> u32 val;
> + bool done = false;
>
> do {
> val = readl(i2cd->regs + I2C_MST_CNTL);
> - if (!(val & I2C_MST_CNTL_CYCLE_TRIGGER))
> - break;
> - if ((val & I2C_MST_CNTL_STATUS) !=
> - I2C_MST_CNTL_STATUS_BUS_BUSY)
> + if (!(val & I2C_MST_CNTL_CYCLE_TRIGGER)
> + || (val & I2C_MST_CNTL_STATUS) !=
> + I2C_MST_CNTL_STATUS_BUS_BUSY) {
Bad formatting. But see below.
> + done = true;
> break;
> + }
> usleep_range(500, 600);
> } while (time_is_after_jiffies(target));
>
> - if (time_is_before_jiffies(target)) {
> + if (!done) {
> dev_err(i2cd->dev, "i2c timeout error %x\n", val);
> return -ETIMEDOUT;
> }
Overall it can use simple tries since you already have sleep inside, but
moreover, you may simple switch to readl_poll_timeout() this entire
loop.
--
With Best Regards,
Andy Shevchenko
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH v2] i2c: nvidia-gpu: Handle timeout correctly in gpu_i2c_check_status()
2020-03-24 13:41 ` Andy Shevchenko
@ 2020-03-24 14:39 ` Kai-Heng Feng
0 siblings, 0 replies; 3+ messages in thread
From: Kai-Heng Feng @ 2020-03-24 14:39 UTC (permalink / raw)
To: Andy Shevchenko
Cc: ajayg, Wolfram Sang,
open list:I2C CONTROLLER DRIVER FOR NVIDIA GPU, open list
Hi Andy,
> On Mar 24, 2020, at 21:41, Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote:
>
> On Tue, Mar 24, 2020 at 09:07:12PM +0800, Kai-Heng Feng wrote:
>> Nvidia card may come with a "phantom" UCSI device, and its driver gets
>> stuck in probe routine, prevents any system PM operations like suspend.
>>
>> When the target time equals to jiffies, it's not included by
>> time_is_before_jiffies(). So let's use a boolean to make sure the
>> operation is done or timeout.
>
> Thank you for an update, my comments below.
>
>> unsigned long target = jiffies + msecs_to_jiffies(1000);
>> u32 val;
>> + bool done = false;
>>
>> do {
>> val = readl(i2cd->regs + I2C_MST_CNTL);
>> - if (!(val & I2C_MST_CNTL_CYCLE_TRIGGER))
>> - break;
>> - if ((val & I2C_MST_CNTL_STATUS) !=
>> - I2C_MST_CNTL_STATUS_BUS_BUSY)
>
>> + if (!(val & I2C_MST_CNTL_CYCLE_TRIGGER)
>> + || (val & I2C_MST_CNTL_STATUS) !=
>> + I2C_MST_CNTL_STATUS_BUS_BUSY) {
>
> Bad formatting. But see below.
>
>> + done = true;
>> break;
>> + }
>> usleep_range(500, 600);
>> } while (time_is_after_jiffies(target));
>>
>> - if (time_is_before_jiffies(target)) {
>> + if (!done) {
>> dev_err(i2cd->dev, "i2c timeout error %x\n", val);
>> return -ETIMEDOUT;
>> }
>
>
> Overall it can use simple tries since you already have sleep inside, but
> moreover, you may simple switch to readl_poll_timeout() this entire
> loop.
Yes that can make the retry loop much simpler.
I'll send a v3 based on your suggestion.
Kai-Heng
>
> --
> With Best Regards,
> Andy Shevchenko
>
>
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2020-03-24 14:39 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-03-24 13:07 [PATCH v2] i2c: nvidia-gpu: Handle timeout correctly in gpu_i2c_check_status() Kai-Heng Feng
2020-03-24 13:41 ` Andy Shevchenko
2020-03-24 14:39 ` Kai-Heng Feng
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox