The intvec handler stores the caches ee in a local variable for use in
processing the intvec. When determining if a syserr is a fatal error or
not, the intvec handler is using the cached version, when it should be
using the current ee read from the device. Currently, the device could
be in the PBL ee as the result of a fatal error, but the cached ee might
be AMSS, which would cause the intvec handler to incorrectly signal a
non-fatal syserr.
Fixes: 3000f85b8f47 ("bus: mhi: core: Add support for basic PM operations")
Signed-off-by: Jeffrey Hugo <[email protected]>
---
drivers/bus/mhi/core/main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/bus/mhi/core/main.c b/drivers/bus/mhi/core/main.c
index 4e0131b..f182736 100644
--- a/drivers/bus/mhi/core/main.c
+++ b/drivers/bus/mhi/core/main.c
@@ -448,7 +448,7 @@ irqreturn_t mhi_intvec_threaded_handler(int irq_number, void *priv)
wake_up_all(&mhi_cntrl->state_event);
/* For fatal errors, we let controller decide next step */
- if (MHI_IN_PBL(ee))
+ if (MHI_IN_PBL(mhi_cntrl->ee))
mhi_cntrl->status_cb(mhi_cntrl, MHI_CB_FATAL_ERROR);
else
mhi_pm_sys_err_handler(mhi_cntrl);
--
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.
On Sat, 13 Feb 2021 at 01:41, Jeffrey Hugo <[email protected]> wrote:
>
> The intvec handler stores the caches ee in a local variable for use in
> processing the intvec. When determining if a syserr is a fatal error or
> not, the intvec handler is using the cached version, when it should be
> using the current ee read from the device. Currently, the device could
> be in the PBL ee as the result of a fatal error, but the cached ee might
> be AMSS, which would cause the intvec handler to incorrectly signal a
> non-fatal syserr.
>
> Fixes: 3000f85b8f47 ("bus: mhi: core: Add support for basic PM operations")
> Signed-off-by: Jeffrey Hugo <[email protected]>
So in that function, 'ee' should be 'previous_ee', right?
Reviewed-by: Loic Poulain <[email protected]>
On Fri, Feb 12, 2021 at 05:40:14PM -0700, Jeffrey Hugo wrote:
> The intvec handler stores the caches ee in a local variable for use in
> processing the intvec. When determining if a syserr is a fatal error or
> not, the intvec handler is using the cached version, when it should be
> using the current ee read from the device. Currently, the device could
> be in the PBL ee as the result of a fatal error, but the cached ee might
> be AMSS, which would cause the intvec handler to incorrectly signal a
> non-fatal syserr.
>
> Fixes: 3000f85b8f47 ("bus: mhi: core: Add support for basic PM operations")
> Signed-off-by: Jeffrey Hugo <[email protected]>
Reviewed-by: Manivannan Sadhasivam <[email protected]>
Thanks,
Mani
> ---
> drivers/bus/mhi/core/main.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/bus/mhi/core/main.c b/drivers/bus/mhi/core/main.c
> index 4e0131b..f182736 100644
> --- a/drivers/bus/mhi/core/main.c
> +++ b/drivers/bus/mhi/core/main.c
> @@ -448,7 +448,7 @@ irqreturn_t mhi_intvec_threaded_handler(int irq_number, void *priv)
> wake_up_all(&mhi_cntrl->state_event);
>
> /* For fatal errors, we let controller decide next step */
> - if (MHI_IN_PBL(ee))
> + if (MHI_IN_PBL(mhi_cntrl->ee))
> mhi_cntrl->status_cb(mhi_cntrl, MHI_CB_FATAL_ERROR);
> else
> mhi_pm_sys_err_handler(mhi_cntrl);
> --
> Qualcomm Technologies, Inc. is a member of the
> Code Aurora Forum, a Linux Foundation Collaborative Project.
>
On 2021-02-15 11:32 PM, Manivannan Sadhasivam wrote:
> On Fri, Feb 12, 2021 at 05:40:14PM -0700, Jeffrey Hugo wrote:
>> The intvec handler stores the caches ee in a local variable for use in
>> processing the intvec. When determining if a syserr is a fatal error
>> or
>> not, the intvec handler is using the cached version, when it should be
>> using the current ee read from the device. Currently, the device
>> could
>> be in the PBL ee as the result of a fatal error, but the cached ee
>> might
>> be AMSS, which would cause the intvec handler to incorrectly signal a
>> non-fatal syserr.
>>
>> Fixes: 3000f85b8f47 ("bus: mhi: core: Add support for basic PM
>> operations")
>> Signed-off-by: Jeffrey Hugo <[email protected]>
>
> Reviewed-by: Manivannan Sadhasivam <[email protected]>
>
> Thanks,
> Mani
>
>> ---
>> drivers/bus/mhi/core/main.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/bus/mhi/core/main.c b/drivers/bus/mhi/core/main.c
>> index 4e0131b..f182736 100644
>> --- a/drivers/bus/mhi/core/main.c
>> +++ b/drivers/bus/mhi/core/main.c
>> @@ -448,7 +448,7 @@ irqreturn_t mhi_intvec_threaded_handler(int
>> irq_number, void *priv)
>> wake_up_all(&mhi_cntrl->state_event);
>>
>> /* For fatal errors, we let controller decide next step */
>> - if (MHI_IN_PBL(ee))
>> + if (MHI_IN_PBL(mhi_cntrl->ee))
>> mhi_cntrl->status_cb(mhi_cntrl, MHI_CB_FATAL_ERROR);
>> else
>> mhi_pm_sys_err_handler(mhi_cntrl);
>> --
>> Qualcomm Technologies, Inc. is a member of the
>> Code Aurora Forum, a Linux Foundation Collaborative Project.
>>
This was being fixed by my patch series for execution environment.
I should just send the most important of my patches as one line fixers
from now on.
Thanks,
Bhaumik
---
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora
Forum,
a Linux Foundation Collaborative Project
On 2/16/2021 10:36 AM, Bhaumik Bhatt wrote:
> On 2021-02-15 11:32 PM, Manivannan Sadhasivam wrote:
>> On Fri, Feb 12, 2021 at 05:40:14PM -0700, Jeffrey Hugo wrote:
>>> The intvec handler stores the caches ee in a local variable for use in
>>> processing the intvec. When determining if a syserr is a fatal error or
>>> not, the intvec handler is using the cached version, when it should be
>>> using the current ee read from the device. Currently, the device could
>>> be in the PBL ee as the result of a fatal error, but the cached ee might
>>> be AMSS, which would cause the intvec handler to incorrectly signal a
>>> non-fatal syserr.
>>>
>>> Fixes: 3000f85b8f47 ("bus: mhi: core: Add support for basic PM
>>> operations")
>>> Signed-off-by: Jeffrey Hugo <[email protected]>
>>
>> Reviewed-by: Manivannan Sadhasivam <[email protected]>
>>
>> Thanks,
>> Mani
>>
>>> ---
>>> drivers/bus/mhi/core/main.c | 2 +-
>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/bus/mhi/core/main.c b/drivers/bus/mhi/core/main.c
>>> index 4e0131b..f182736 100644
>>> --- a/drivers/bus/mhi/core/main.c
>>> +++ b/drivers/bus/mhi/core/main.c
>>> @@ -448,7 +448,7 @@ irqreturn_t mhi_intvec_threaded_handler(int
>>> irq_number, void *priv)
>>> wake_up_all(&mhi_cntrl->state_event);
>>>
>>> /* For fatal errors, we let controller decide next step */
>>> - if (MHI_IN_PBL(ee))
>>> + if (MHI_IN_PBL(mhi_cntrl->ee))
>>> mhi_cntrl->status_cb(mhi_cntrl, MHI_CB_FATAL_ERROR);
>>> else
>>> mhi_pm_sys_err_handler(mhi_cntrl);
>>> --
>>> Qualcomm Technologies, Inc. is a member of the
>>> Code Aurora Forum, a Linux Foundation Collaborative Project.
>>>
>
> This was being fixed by my patch series for execution environment.
>
> I should just send the most important of my patches as one line fixers
> from now on.
Its been an issue for some 9 months now. I really don't care how it
gets fixed, but it seems to be past time for it to be fixed.
Sending simple little fixes on their own seems to accelerate the review
process of those fixes, and doesn't hold them back if some "dependency"
in a series has comments which need to be fixed. It is a judgement call
to determine if a set of fixes should be sent together or not, but I
personally try to send individual patches when I can.
--
Jeffrey Hugo
Qualcomm Technologies, Inc. is a member of the
Code Aurora Forum, a Linux Foundation Collaborative Project.