On 05.02.2022 14:53, Greg Kroah-Hartman wrote:
> On Sat, Feb 05, 2022 at 02:40:37PM +0300, Alexey Khoroshilov wrote:
>> On 04.02.2022 12:20, Greg Kroah-Hartman wrote:
>>> From: Maxime Ripard <maxime@xxxxxxxxxx>
>>>
>>> Commit 20b0dfa86bef0e80b41b0e5ac38b92f23b6f27f9 upstream.
>>>
>>> The original commit depended on a rework commit (724fc856c09e ("drm/vc4:
>>> hdmi: Split the CEC disable / enable functions in two")) that
>>> (rightfully) didn't reach stable.
>>>
>>> However, probably because the context changed, when the patch was
>>> applied to stable the pm_runtime_put called got moved to the end of the
>>> vc4_hdmi_cec_adap_enable function (that would have become
>>> vc4_hdmi_cec_disable with the rework) to vc4_hdmi_cec_init.
>>>
>>> This means that at probe time, we now drop our reference to the clocks
>>> and power domains and thus end up with a CPU hang when the CPU tries to
>>> access registers.
>>>
>>> The call to pm_runtime_resume_and_get() is also problematic since the
>>> .adap_enable CEC hook is called both to enable and to disable the
>>> controller. That means that we'll now call pm_runtime_resume_and_get()
>>> at disable time as well, messing with the reference counting.
>>>
>>> The behaviour we should have though would be to have
>>> pm_runtime_resume_and_get() called when the CEC controller is enabled,
>>> and pm_runtime_put when it's disabled.
>>>
>>> We need to move things around a bit to behave that way, but it aligns
>>> stable with upstream.
>>>
>>> Cc: <stable@xxxxxxxxxxxxxxx> # 5.10.x
>>> Cc: <stable@xxxxxxxxxxxxxxx> # 5.15.x
>>> Cc: <stable@xxxxxxxxxxxxxxx> # 5.16.x
>>> Reported-by: Michael Stapelberg <michael+drm@xxxxxxxxxxxxx>
>>> Signed-off-by: Maxime Ripard <maxime@xxxxxxxxxx>
>>> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
>>> ---
>>> drivers/gpu/drm/vc4/vc4_hdmi.c | 27 ++++++++++++++-------------
>>> 1 file changed, 14 insertions(+), 13 deletions(-)
>>>
>>> --- a/drivers/gpu/drm/vc4/vc4_hdmi.c
>>> +++ b/drivers/gpu/drm/vc4/vc4_hdmi.c
>>> @@ -1402,18 +1402,18 @@ static int vc4_hdmi_cec_adap_enable(stru
>>> u32 val;
>>> int ret;
>>>
>>> - ret = pm_runtime_resume_and_get(&vc4_hdmi->pdev->dev);
>>> - if (ret)
>>> - return ret;
>>> -
>>> - val = HDMI_READ(HDMI_CEC_CNTRL_5);
>>> - val &= ~(VC4_HDMI_CEC_TX_SW_RESET | VC4_HDMI_CEC_RX_SW_RESET |
>>> - VC4_HDMI_CEC_CNT_TO_4700_US_MASK |
>>> - VC4_HDMI_CEC_CNT_TO_4500_US_MASK);
>>> - val |= ((4700 / usecs) << VC4_HDMI_CEC_CNT_TO_4700_US_SHIFT) |
>>> - ((4500 / usecs) << VC4_HDMI_CEC_CNT_TO_4500_US_SHIFT);
>>> -
>>> if (enable) {
>>> + ret = pm_runtime_resume_and_get(&vc4_hdmi->pdev->dev);
>>> + if (ret)
>>> + return ret;
>>> +
>>> + val = HDMI_READ(HDMI_CEC_CNTRL_5);
>>> + val &= ~(VC4_HDMI_CEC_TX_SW_RESET | VC4_HDMI_CEC_RX_SW_RESET |
>>> + VC4_HDMI_CEC_CNT_TO_4700_US_MASK |
>>> + VC4_HDMI_CEC_CNT_TO_4500_US_MASK);
>>> + val |= ((4700 / usecs) << VC4_HDMI_CEC_CNT_TO_4700_US_SHIFT) |
>>> + ((4500 / usecs) << VC4_HDMI_CEC_CNT_TO_4500_US_SHIFT);
>>> +
>>> HDMI_WRITE(HDMI_CEC_CNTRL_5, val |
>>> VC4_HDMI_CEC_TX_SW_RESET | VC4_HDMI_CEC_RX_SW_RESET);
>>> HDMI_WRITE(HDMI_CEC_CNTRL_5, val);
>>> @@ -1439,7 +1439,10 @@ static int vc4_hdmi_cec_adap_enable(stru
>>> HDMI_WRITE(HDMI_CEC_CPU_MASK_SET, VC4_HDMI_CPU_CEC);
>>> HDMI_WRITE(
, val |
>>> VC4_HDMI_CEC_TX_SW_RESET | VC4_HDMI_CEC_RX_SW_RESET);
>>> +
>>> + pm_runtime_put(&vc4_hdmi->pdev->dev);
>>> }
>>> +
>>> return 0;
>>> }
>>>
>>> @@ -1531,8 +1534,6 @@ static int vc4_hdmi_cec_init(struct vc4_
>>> if (ret < 0)
>>> goto err_delete_cec_adap;
>>>
>>> - pm_runtime_put(&vc4_hdmi->pdev->dev);
>>> -
>>> return 0;
>>>
>>> err_delete_cec_adap:
>>>
>>>
>>
>> The patch has moved initialization of val local variable into if
>> (enable) branch. But the variable is used in in the else branch as well.
>> As a result we write of its initialized value here:
>>
>> HDMI_WRITE(HDMI_CEC_CNTRL_5, val |
>> VC4_HDMI_CEC_TX_SW_RESET | VC4_HDMI_CEC_RX_SW_RESET);
>>
>> Found by Linux Verification Center (linuxtesting.org) with SVACE.
>>
>> static
>> int vc4_hdmi_cec_adap_enable(struct cec_adapter *adap, bool enable)
>> {
>> struct vc4_hdmi *vc4_hdmi = cec_get_drvdata(adap);
>> /* clock period in microseconds */
>> const u32 usecs = 1000000 / CEC_CLOCK_FREQ;
>> u32 val;
>> int ret;
>>
>> if (enable) {
>> ret = pm_runtime_resume_and_get(&vc4_hdmi->pdev->dev);
>> if (ret)
>> return ret;
>>
>> val = HDMI_READ(HDMI_CEC_CNTRL_5);
>> .....
>>
>> } else {
>> HDMI_WRITE(HDMI_CEC_CPU_MASK_SET, VC4_HDMI_CPU_CEC);
>> HDMI_WRITE(HDMI_CEC_CNTRL_5, val | <------------------ UNINIT VALUE
>> VC4_HDMI_CEC_TX_SW_RESET | VC4_HDMI_CEC_RX_SW_RESET);
>>
>> pm_runtime_put(&vc4_hdmi->pdev->dev);
>> }
>>
>> return 0;
>> }
>
> So what does this mean? That this backport is incorrect and should be
> dropped? Or that the original commit was wrong? Or something else?
>
> confused,
>
> greg k-h
>
As far as I can see, the upstream commit was correct and the problem is
in the backport. I would suggest to fix it like this:
} else {
HDMI_WRITE(HDMI_CEC_CPU_MASK_SET, VC4_HDMI_CPU_CEC);
- HDMI_WRITE(HDMI_CEC_CNTRL_5, val
+ HDMI_WRITE(HDMI_CEC_CNTRL_5, HDMI_READ(HDMI_CEC_CNTRL_5)
VC4_HDMI_CEC_TX_SW_RESET | VC4_HDMI_CEC_RX_SW_RESET);
pm_runtime_put(&vc4_hdmi->pdev->dev);
}
But I would left it to comment someone who knows this code better.
--
Best regards,
Alexey Khoroshilov
Linux Verification Center, ISPRAS
