mbox series

[RFC,PULL] ath11k: move WCN6750 firmware to the device-specific subdir

Message ID 20241011061606.1612918-1-dmitry.baryshkov@linaro.org (mailing list archive)
State Not Applicable
Headers show
Series [RFC,PULL] ath11k: move WCN6750 firmware to the device-specific subdir | expand

Pull-request

https://gitlab.com/lumag/linux-firmware wcn6750-device-specific

Message

Dmitry Baryshkov Oct. 11, 2024, 6:16 a.m. UTC
The firmware under ath11k/WCN6750 can not be utilized on other devices using
WCN6750 to provide WiFi (e.g. on the Qualcomm RB3 gen2 board).  This is an
attempt to propose a solution by moving the file into device-specifig subdir.
Other platforms can place firmware files in other subdirs, depending on the
SoC.

Note, while performing the move I also propose squashing the firmware file.
This is the practice recommended by the msm platform maintainers for the DSP
firmware (not to mention that it also saves us from the extra symlinks). If
Kalle, ath11k maintainer, objects, I will keep the firmware in the split form.

For now this is an RFC in order to gather opinion form respective platform and
driver maintainers.

----------------------------------------------------------------

The following changes since commit d55315295cccf8ce5dc288a74d75cf09b8063cfc:

  Merge branch 'superm1/qci-patches' into 'main' (2024-10-10 19:11:13 +0000)

are available in the Git repository at:

  https://gitlab.com/lumag/linux-firmware wcn6750-device-specific

for you to fetch changes up to abe8c109a2b4c422b627335d124473734c8790d8:

  ath11k: move WCN6750 firmware to the device-specific subdir (2024-10-10 22:26:20 +0300)

----------------------------------------------------------------
Dmitry Baryshkov (1):
      ath11k: move WCN6750 firmware to the device-specific subdir

 WHENCE                                             |  12 ++----------
 ath11k/WCN6750/hw1.0/{wpss.b04 => sc7280/wpss.mbn} | Bin 5819999 -> 7463728 bytes
 ath11k/WCN6750/hw1.0/wpss.b00                      | Bin 340 -> 0 bytes
 ath11k/WCN6750/hw1.0/wpss.b01                      | Bin 6848 -> 0 bytes
 ath11k/WCN6750/hw1.0/wpss.b02                      | Bin 10300 -> 0 bytes
 ath11k/WCN6750/hw1.0/wpss.b03                      | Bin 4096 -> 0 bytes
 ath11k/WCN6750/hw1.0/wpss.b05                      | Bin 164332 -> 0 bytes
 ath11k/WCN6750/hw1.0/wpss.b06                      | Bin 266684 -> 0 bytes
 ath11k/WCN6750/hw1.0/wpss.b07                      | Bin 1176368 -> 0 bytes
 ath11k/WCN6750/hw1.0/wpss.b08                      |   0
 ath11k/WCN6750/hw1.0/wpss.mdt                      | Bin 7188 -> 0 bytes
 11 files changed, 2 insertions(+), 10 deletions(-)
 rename ath11k/WCN6750/hw1.0/{wpss.b04 => sc7280/wpss.mbn} (77%)
 delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b00
 delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b01
 delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b02
 delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b03
 delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b05
 delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b06
 delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b07
 delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b08
 delete mode 100644 ath11k/WCN6750/hw1.0/wpss.mdt

Comments

Jeff Johnson Oct. 18, 2024, 10:16 p.m. UTC | #1
On 10/10/2024 11:16 PM, Dmitry Baryshkov wrote:
> The firmware under ath11k/WCN6750 can not be utilized on other devices using
> WCN6750 to provide WiFi (e.g. on the Qualcomm RB3 gen2 board).  This is an
> attempt to propose a solution by moving the file into device-specifig subdir.
> Other platforms can place firmware files in other subdirs, depending on the
> SoC.
> 
> Note, while performing the move I also propose squashing the firmware file.
> This is the practice recommended by the msm platform maintainers for the DSP
> firmware (not to mention that it also saves us from the extra symlinks). If
> Kalle, ath11k maintainer, objects, I will keep the firmware in the split form.
> 
> For now this is an RFC in order to gather opinion form respective platform and
> driver maintainers.
> 
> ----------------------------------------------------------------
> 
> The following changes since commit d55315295cccf8ce5dc288a74d75cf09b8063cfc:
> 
>   Merge branch 'superm1/qci-patches' into 'main' (2024-10-10 19:11:13 +0000)
> 
> are available in the Git repository at:
> 
>   https://gitlab.com/lumag/linux-firmware wcn6750-device-specific
> 
> for you to fetch changes up to abe8c109a2b4c422b627335d124473734c8790d8:
> 
>   ath11k: move WCN6750 firmware to the device-specific subdir (2024-10-10 22:26:20 +0300)
> 
> ----------------------------------------------------------------
> Dmitry Baryshkov (1):
>       ath11k: move WCN6750 firmware to the device-specific subdir
> 
>  WHENCE                                             |  12 ++----------
>  ath11k/WCN6750/hw1.0/{wpss.b04 => sc7280/wpss.mbn} | Bin 5819999 -> 7463728 bytes
>  ath11k/WCN6750/hw1.0/wpss.b00                      | Bin 340 -> 0 bytes
>  ath11k/WCN6750/hw1.0/wpss.b01                      | Bin 6848 -> 0 bytes
>  ath11k/WCN6750/hw1.0/wpss.b02                      | Bin 10300 -> 0 bytes
>  ath11k/WCN6750/hw1.0/wpss.b03                      | Bin 4096 -> 0 bytes
>  ath11k/WCN6750/hw1.0/wpss.b05                      | Bin 164332 -> 0 bytes
>  ath11k/WCN6750/hw1.0/wpss.b06                      | Bin 266684 -> 0 bytes
>  ath11k/WCN6750/hw1.0/wpss.b07                      | Bin 1176368 -> 0 bytes
>  ath11k/WCN6750/hw1.0/wpss.b08                      |   0
>  ath11k/WCN6750/hw1.0/wpss.mdt                      | Bin 7188 -> 0 bytes
>  11 files changed, 2 insertions(+), 10 deletions(-)
>  rename ath11k/WCN6750/hw1.0/{wpss.b04 => sc7280/wpss.mbn} (77%)
>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b00
>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b01
>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b02
>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b03
>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b05
>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b06
>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b07
>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b08
>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.mdt

I'm waiting for Kalle to respond, but at a minimum I suspect our internal
firmware tooling would be impacted by this change.

Kalle, do you have concerns or suggestions on how to handle this issue?

/jeff
Kalle Valo Oct. 21, 2024, 6:22 p.m. UTC | #2
Jeff Johnson <quic_jjohnson@quicinc.com> writes:

> On 10/10/2024 11:16 PM, Dmitry Baryshkov wrote:
>
>> The firmware under ath11k/WCN6750 can not be utilized on other devices using
>> WCN6750 to provide WiFi (e.g. on the Qualcomm RB3 gen2 board).  This is an
>> attempt to propose a solution by moving the file into device-specifig subdir.
>> Other platforms can place firmware files in other subdirs, depending on the
>> SoC.
>> 
>> Note, while performing the move I also propose squashing the firmware
>> file.

I assume you mean squashing the files to wpss.mbn?

>> This is the practice recommended by the msm platform maintainers for the DSP
>> firmware (not to mention that it also saves us from the extra symlinks). If
>> Kalle, ath11k maintainer, objects, I will keep the firmware in the split form.
>> 
>> For now this is an RFC in order to gather opinion form respective platform and
>> driver maintainers.
>> 
>> ----------------------------------------------------------------
>> 
>> The following changes since commit d55315295cccf8ce5dc288a74d75cf09b8063cfc:
>> 
>>   Merge branch 'superm1/qci-patches' into 'main' (2024-10-10 19:11:13 +0000)
>> 
>> are available in the Git repository at:
>> 
>>   https://gitlab.com/lumag/linux-firmware wcn6750-device-specific
>> 
>> for you to fetch changes up to abe8c109a2b4c422b627335d124473734c8790d8:
>> 
>>   ath11k: move WCN6750 firmware to the device-specific subdir (2024-10-10 22:26:20 +0300)
>> 
>> ----------------------------------------------------------------
>> Dmitry Baryshkov (1):
>>       ath11k: move WCN6750 firmware to the device-specific subdir
>> 
>>  WHENCE                                             |  12 ++----------
>>  ath11k/WCN6750/hw1.0/{wpss.b04 => sc7280/wpss.mbn} | Bin 5819999 -> 7463728 bytes
>>  ath11k/WCN6750/hw1.0/wpss.b00                      | Bin 340 -> 0 bytes
>>  ath11k/WCN6750/hw1.0/wpss.b01                      | Bin 6848 -> 0 bytes
>>  ath11k/WCN6750/hw1.0/wpss.b02                      | Bin 10300 -> 0 bytes
>>  ath11k/WCN6750/hw1.0/wpss.b03                      | Bin 4096 -> 0 bytes
>>  ath11k/WCN6750/hw1.0/wpss.b05                      | Bin 164332 -> 0 bytes
>>  ath11k/WCN6750/hw1.0/wpss.b06                      | Bin 266684 -> 0 bytes
>>  ath11k/WCN6750/hw1.0/wpss.b07                      | Bin 1176368 -> 0 bytes
>>  ath11k/WCN6750/hw1.0/wpss.b08                      |   0
>>  ath11k/WCN6750/hw1.0/wpss.mdt                      | Bin 7188 -> 0 bytes
>>  11 files changed, 2 insertions(+), 10 deletions(-)
>>  rename ath11k/WCN6750/hw1.0/{wpss.b04 => sc7280/wpss.mbn} (77%)

So this follow basically the same directory hierarchy as you Dmitry
implemented for ath10k in commit 5abf259772df ("wifi: ath10k: support
board-specific firmware overrides"), right? Just making sure we are on
the same page. I think ath10k/ath11k/ath12k drivers should follow the
same design as much as possible.

>>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b00
>>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b01
>>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b02
>>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b03
>>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b05
>>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b06
>>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b07
>>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b08
>>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.mdt
>
> I'm waiting for Kalle to respond, but at a minimum I suspect our internal
> firmware tooling would be impacted by this change.

Yeah, most likely we need to update our tools. But we would have to do
it anyway so I'm not concerned about that.

> Kalle, do you have concerns or suggestions on how to handle this issue?

Looks good to me.

Dmitry, if you have the time please take a look at this:

https://patchwork.kernel.org/project/linux-wireless/cover/20241001033053.2084360-1-quic_miaoqing@quicinc.com/

Based on a quick look it seems to do things differently and not really a
fan of that.
Dmitry Baryshkov Oct. 21, 2024, 6:34 p.m. UTC | #3
On Mon, 21 Oct 2024 at 21:22, Kalle Valo <kvalo@kernel.org> wrote:
>
> Jeff Johnson <quic_jjohnson@quicinc.com> writes:
>
> > On 10/10/2024 11:16 PM, Dmitry Baryshkov wrote:
> >
> >> The firmware under ath11k/WCN6750 can not be utilized on other devices using
> >> WCN6750 to provide WiFi (e.g. on the Qualcomm RB3 gen2 board).  This is an
> >> attempt to propose a solution by moving the file into device-specifig subdir.
> >> Other platforms can place firmware files in other subdirs, depending on the
> >> SoC.
> >>
> >> Note, while performing the move I also propose squashing the firmware
> >> file.
>
> I assume you mean squashing the files to wpss.mbn?
>
> >> This is the practice recommended by the msm platform maintainers for the DSP
> >> firmware (not to mention that it also saves us from the extra symlinks). If
> >> Kalle, ath11k maintainer, objects, I will keep the firmware in the split form.
> >>
> >> For now this is an RFC in order to gather opinion form respective platform and
> >> driver maintainers.
> >>
> >> ----------------------------------------------------------------
> >>
> >> The following changes since commit d55315295cccf8ce5dc288a74d75cf09b8063cfc:
> >>
> >>   Merge branch 'superm1/qci-patches' into 'main' (2024-10-10 19:11:13 +0000)
> >>
> >> are available in the Git repository at:
> >>
> >>   https://gitlab.com/lumag/linux-firmware wcn6750-device-specific
> >>
> >> for you to fetch changes up to abe8c109a2b4c422b627335d124473734c8790d8:
> >>
> >>   ath11k: move WCN6750 firmware to the device-specific subdir (2024-10-10 22:26:20 +0300)
> >>
> >> ----------------------------------------------------------------
> >> Dmitry Baryshkov (1):
> >>       ath11k: move WCN6750 firmware to the device-specific subdir
> >>
> >>  WHENCE                                             |  12 ++----------
> >>  ath11k/WCN6750/hw1.0/{wpss.b04 => sc7280/wpss.mbn} | Bin 5819999 -> 7463728 bytes
> >>  ath11k/WCN6750/hw1.0/wpss.b00                      | Bin 340 -> 0 bytes
> >>  ath11k/WCN6750/hw1.0/wpss.b01                      | Bin 6848 -> 0 bytes
> >>  ath11k/WCN6750/hw1.0/wpss.b02                      | Bin 10300 -> 0 bytes
> >>  ath11k/WCN6750/hw1.0/wpss.b03                      | Bin 4096 -> 0 bytes
> >>  ath11k/WCN6750/hw1.0/wpss.b05                      | Bin 164332 -> 0 bytes
> >>  ath11k/WCN6750/hw1.0/wpss.b06                      | Bin 266684 -> 0 bytes
> >>  ath11k/WCN6750/hw1.0/wpss.b07                      | Bin 1176368 -> 0 bytes
> >>  ath11k/WCN6750/hw1.0/wpss.b08                      |   0
> >>  ath11k/WCN6750/hw1.0/wpss.mdt                      | Bin 7188 -> 0 bytes
> >>  11 files changed, 2 insertions(+), 10 deletions(-)
> >>  rename ath11k/WCN6750/hw1.0/{wpss.b04 => sc7280/wpss.mbn} (77%)
>
> So this follow basically the same directory hierarchy as you Dmitry
> implemented for ath10k in commit 5abf259772df ("wifi: ath10k: support
> board-specific firmware overrides"), right? Just making sure we are on
> the same page. I think ath10k/ath11k/ath12k drivers should follow the
> same design as much as possible.

Yes, that was my goal. Having subdirs under the main dir.

>
> >>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b00
> >>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b01
> >>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b02
> >>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b03
> >>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b05
> >>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b06
> >>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b07
> >>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.b08
> >>  delete mode 100644 ath11k/WCN6750/hw1.0/wpss.mdt
> >
> > I'm waiting for Kalle to respond, but at a minimum I suspect our internal
> > firmware tooling would be impacted by this change.
>
> Yeah, most likely we need to update our tools. But we would have to do
> it anyway so I'm not concerned about that.

Ok. If there are no concerns from your side, could you possibly
respond with Acked-by? Then we can unblock and merge this.

> > Kalle, do you have concerns or suggestions on how to handle this issue?
>
> Looks good to me.
>
> Dmitry, if you have the time please take a look at this:
>
> https://patchwork.kernel.org/project/linux-wireless/cover/20241001033053.2084360-1-quic_miaoqing@quicinc.com/
>
> Based on a quick look it seems to do things differently and not really a
> fan of that.

Yes, I'd prefer a subdir rather than this implementation. I'll respond
either today or tomorrow.