On Tue, May 14, 2024 at 11:43 PM Jeff Johnson <[email protected]> wrote:
>
> On 5/11/2024 3:49 AM, Dmitry Baryshkov wrote:
> > If ath10k_snoc is built-in, while Qualcomm remoteprocs are built as
> > modules, compilation fails with:
> >
> > /usr/bin/aarch64-linux-gnu-ld: drivers/net/wireless/ath/ath10k/snoc.o: in function `ath10k_modem_init':
> > drivers/net/wireless/ath/ath10k/snoc.c:1534: undefined reference to `qcom_register_ssr_notifier'
> > /usr/bin/aarch64-linux-gnu-ld: drivers/net/wireless/ath/ath10k/snoc.o: in function `ath10k_modem_deinit':
> > drivers/net/wireless/ath/ath10k/snoc.c:1551: undefined reference to `qcom_unregister_ssr_notifier'
> >
> > Add corresponding dependency to ATH10K_SNOC Kconfig entry so that it's
> > built as module if QCOM_RPROC_COMMON is built as module too.
> >
> > Fixes: 747ff7d3d742 ("ath10k: Don't always treat modem stop events as crashes")
> > Cc: [email protected]
> > Signed-off-by: Dmitry Baryshkov <[email protected]>
> > ---
> > drivers/net/wireless/ath/ath10k/Kconfig | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/net/wireless/ath/ath10k/Kconfig b/drivers/net/wireless/ath/ath10k/Kconfig
> > index e6ea884cafc1..4f385f4a8cef 100644
> > --- a/drivers/net/wireless/ath/ath10k/Kconfig
> > +++ b/drivers/net/wireless/ath/ath10k/Kconfig
> > @@ -45,6 +45,7 @@ config ATH10K_SNOC
> > depends on ATH10K
> > depends on ARCH_QCOM || COMPILE_TEST
> > depends on QCOM_SMEM
> > + depends on QCOM_RPROC_COMMON || QCOM_RPROC_COMMON=n
> > select QCOM_SCM
> > select QCOM_QMI_HELPERS
> > help
> >
> > ---
> > base-commit: 75fa778d74b786a1608d55d655d42b480a6fa8bd
> > change-id: 20240511-ath10k-snoc-dep-862a9da2e6bb
>
> I see how this fixes the problem, but this doesn't seem like an ideal
> solution. The fact that the *_ssr_notifier() functions are correctly protected
> by conditional compilation ideally should mean that clients don't need to call
> call out this as a dependency. Otherwise, it would mean we'd need to do this
> for all feature flags.
It depends on if qcom_common.c is optional for ath10k_snoc.
If it is optional, this patch is correct.
If it is mandatory
depends on QCOM_RPROC_COMMON
>
> +linux-kbuild just to make sure there isn't a better approach.
Commit 28d49e171676afb7df7f47798391364af9abed7f suggested
depends on BAR || !BAR
but
depends on BAR || BAR=n
works equivalently.
>
> /jeff
>
>
--
Best Regards
Masahiro Yamada
On 5/14/2024 10:45 AM, Masahiro Yamada wrote:
> On Tue, May 14, 2024 at 11:43 PM Jeff Johnson <[email protected]> wrote:
>>
>> On 5/11/2024 3:49 AM, Dmitry Baryshkov wrote:
>>> If ath10k_snoc is built-in, while Qualcomm remoteprocs are built as
>>> modules, compilation fails with:
>>>
>>> /usr/bin/aarch64-linux-gnu-ld: drivers/net/wireless/ath/ath10k/snoc.o: in function `ath10k_modem_init':
>>> drivers/net/wireless/ath/ath10k/snoc.c:1534: undefined reference to `qcom_register_ssr_notifier'
>>> /usr/bin/aarch64-linux-gnu-ld: drivers/net/wireless/ath/ath10k/snoc.o: in function `ath10k_modem_deinit':
>>> drivers/net/wireless/ath/ath10k/snoc.c:1551: undefined reference to `qcom_unregister_ssr_notifier'
>>>
>>> Add corresponding dependency to ATH10K_SNOC Kconfig entry so that it's
>>> built as module if QCOM_RPROC_COMMON is built as module too.
>>>
>>> Fixes: 747ff7d3d742 ("ath10k: Don't always treat modem stop events as crashes")
>>> Cc: [email protected]
>>> Signed-off-by: Dmitry Baryshkov <[email protected]>
>>> ---
>>> drivers/net/wireless/ath/ath10k/Kconfig | 1 +
>>> 1 file changed, 1 insertion(+)
>>>
>>> diff --git a/drivers/net/wireless/ath/ath10k/Kconfig b/drivers/net/wireless/ath/ath10k/Kconfig
>>> index e6ea884cafc1..4f385f4a8cef 100644
>>> --- a/drivers/net/wireless/ath/ath10k/Kconfig
>>> +++ b/drivers/net/wireless/ath/ath10k/Kconfig
>>> @@ -45,6 +45,7 @@ config ATH10K_SNOC
>>> depends on ATH10K
>>> depends on ARCH_QCOM || COMPILE_TEST
>>> depends on QCOM_SMEM
>>> + depends on QCOM_RPROC_COMMON || QCOM_RPROC_COMMON=n
>>> select QCOM_SCM
>>> select QCOM_QMI_HELPERS
>>> help
>>>
>>> ---
>>> base-commit: 75fa778d74b786a1608d55d655d42b480a6fa8bd
>>> change-id: 20240511-ath10k-snoc-dep-862a9da2e6bb
>>
>> I see how this fixes the problem, but this doesn't seem like an ideal
>> solution. The fact that the *_ssr_notifier() functions are correctly protected
>> by conditional compilation ideally should mean that clients don't need to call
>> call out this as a dependency. Otherwise, it would mean we'd need to do this
>> for all feature flags.
>
>
> It depends on if qcom_common.c is optional for ath10k_snoc.
>
> If it is optional, this patch is correct.
At least from a build perspective it is optional
>
>
> If it is mandatory
> depends on QCOM_RPROC_COMMON
>
>
>>
>> +linux-kbuild just to make sure there isn't a better approach.
>
>
> Commit 28d49e171676afb7df7f47798391364af9abed7f suggested
>
> depends on BAR || !BAR
>
> but
>
> depends on BAR || BAR=n
>
> works equivalently.
Thanks for the information.
I'll let Kalle decide if he wants to change this in 'pending' to:
depends on QCOM_RPROC_COMMON || !QCOM_RPROC_COMMON
Acked-by: Jeff Johnson <[email protected]>
On Wed, May 15, 2024 at 3:14 AM Jeff Johnson <[email protected]> wrote:
>
> On 5/14/2024 10:45 AM, Masahiro Yamada wrote:
> > On Tue, May 14, 2024 at 11:43 PM Jeff Johnson <[email protected]> wrote:
> >>
> >> On 5/11/2024 3:49 AM, Dmitry Baryshkov wrote:
> >>> If ath10k_snoc is built-in, while Qualcomm remoteprocs are built as
> >>> modules, compilation fails with:
> >>>
> >>> /usr/bin/aarch64-linux-gnu-ld: drivers/net/wireless/ath/ath10k/snoc.o: in function `ath10k_modem_init':
> >>> drivers/net/wireless/ath/ath10k/snoc.c:1534: undefined reference to `qcom_register_ssr_notifier'
> >>> /usr/bin/aarch64-linux-gnu-ld: drivers/net/wireless/ath/ath10k/snoc.o: in function `ath10k_modem_deinit':
> >>> drivers/net/wireless/ath/ath10k/snoc.c:1551: undefined reference to `qcom_unregister_ssr_notifier'
> >>>
> >>> Add corresponding dependency to ATH10K_SNOC Kconfig entry so that it's
> >>> built as module if QCOM_RPROC_COMMON is built as module too.
> >>>
> >>> Fixes: 747ff7d3d742 ("ath10k: Don't always treat modem stop events as crashes")
> >>> Cc: [email protected]
> >>> Signed-off-by: Dmitry Baryshkov <[email protected]>
> >>> ---
> >>> drivers/net/wireless/ath/ath10k/Kconfig | 1 +
> >>> 1 file changed, 1 insertion(+)
> >>>
> >>> diff --git a/drivers/net/wireless/ath/ath10k/Kconfig b/drivers/net/wireless/ath/ath10k/Kconfig
> >>> index e6ea884cafc1..4f385f4a8cef 100644
> >>> --- a/drivers/net/wireless/ath/ath10k/Kconfig
> >>> +++ b/drivers/net/wireless/ath/ath10k/Kconfig
> >>> @@ -45,6 +45,7 @@ config ATH10K_SNOC
> >>> depends on ATH10K
> >>> depends on ARCH_QCOM || COMPILE_TEST
> >>> depends on QCOM_SMEM
> >>> + depends on QCOM_RPROC_COMMON || QCOM_RPROC_COMMON=n
> >>> select QCOM_SCM
> >>> select QCOM_QMI_HELPERS
> >>> help
> >>>
> >>> ---
> >>> base-commit: 75fa778d74b786a1608d55d655d42b480a6fa8bd
> >>> change-id: 20240511-ath10k-snoc-dep-862a9da2e6bb
> >>
> >> I see how this fixes the problem, but this doesn't seem like an ideal
> >> solution. The fact that the *_ssr_notifier() functions are correctly protected
> >> by conditional compilation ideally should mean that clients don't need to call
> >> call out this as a dependency. Otherwise, it would mean we'd need to do this
> >> for all feature flags.
> >
> >
> > It depends on if qcom_common.c is optional for ath10k_snoc.
> >
> > If it is optional, this patch is correct.
>
> At least from a build perspective it is optional
You need to make a decision from a run-time perspective
(that is, whether you are fine with having
ar_snoc->notifier == NULL)
If you are able to build the kernel successfully
but you get a run-time error (e.g. NULL pointer dereference),
it is even worse because run-time debugging is generally more
difficult than compile-time debugging.
--
Best Regards
Masahiro Yamada
On 5/14/2024 9:10 PM, Masahiro Yamada wrote:
> On Wed, May 15, 2024 at 3:14 AM Jeff Johnson <[email protected]> wrote:
>>
>> On 5/14/2024 10:45 AM, Masahiro Yamada wrote:
>>> On Tue, May 14, 2024 at 11:43 PM Jeff Johnson <[email protected]> wrote:
>>>>
>>>> On 5/11/2024 3:49 AM, Dmitry Baryshkov wrote:
>>>>> If ath10k_snoc is built-in, while Qualcomm remoteprocs are built as
>>>>> modules, compilation fails with:
>>>>>
>>>>> /usr/bin/aarch64-linux-gnu-ld: drivers/net/wireless/ath/ath10k/snoc.o: in function `ath10k_modem_init':
>>>>> drivers/net/wireless/ath/ath10k/snoc.c:1534: undefined reference to `qcom_register_ssr_notifier'
>>>>> /usr/bin/aarch64-linux-gnu-ld: drivers/net/wireless/ath/ath10k/snoc.o: in function `ath10k_modem_deinit':
>>>>> drivers/net/wireless/ath/ath10k/snoc.c:1551: undefined reference to `qcom_unregister_ssr_notifier'
>>>>>
>>>>> Add corresponding dependency to ATH10K_SNOC Kconfig entry so that it's
>>>>> built as module if QCOM_RPROC_COMMON is built as module too.
>>>>>
>>>>> Fixes: 747ff7d3d742 ("ath10k: Don't always treat modem stop events as crashes")
>>>>> Cc: [email protected]
>>>>> Signed-off-by: Dmitry Baryshkov <[email protected]>
>>>>> ---
>>>>> drivers/net/wireless/ath/ath10k/Kconfig | 1 +
>>>>> 1 file changed, 1 insertion(+)
>>>>>
>>>>> diff --git a/drivers/net/wireless/ath/ath10k/Kconfig b/drivers/net/wireless/ath/ath10k/Kconfig
>>>>> index e6ea884cafc1..4f385f4a8cef 100644
>>>>> --- a/drivers/net/wireless/ath/ath10k/Kconfig
>>>>> +++ b/drivers/net/wireless/ath/ath10k/Kconfig
>>>>> @@ -45,6 +45,7 @@ config ATH10K_SNOC
>>>>> depends on ATH10K
>>>>> depends on ARCH_QCOM || COMPILE_TEST
>>>>> depends on QCOM_SMEM
>>>>> + depends on QCOM_RPROC_COMMON || QCOM_RPROC_COMMON=n
>>>>> select QCOM_SCM
>>>>> select QCOM_QMI_HELPERS
>>>>> help
>>>>>
>>>>> ---
>>>>> base-commit: 75fa778d74b786a1608d55d655d42b480a6fa8bd
>>>>> change-id: 20240511-ath10k-snoc-dep-862a9da2e6bb
>>>>
>>>> I see how this fixes the problem, but this doesn't seem like an ideal
>>>> solution. The fact that the *_ssr_notifier() functions are correctly protected
>>>> by conditional compilation ideally should mean that clients don't need to call
>>>> call out this as a dependency. Otherwise, it would mean we'd need to do this
>>>> for all feature flags.
>>>
>>>
>>> It depends on if qcom_common.c is optional for ath10k_snoc.
>>>
>>> If it is optional, this patch is correct.
>>
>> At least from a build perspective it is optional
>
>
>
> You need to make a decision from a run-time perspective
> (that is, whether you are fine with having
> ar_snoc->notifier == NULL)
>
>
> If you are able to build the kernel successfully
> but you get a run-time error (e.g. NULL pointer dereference),
> it is even worse because run-time debugging is generally more
> difficult than compile-time debugging.
The current patch maintains the existing logic, so that is my preference.