mbox series

[v3,0/3] spi:amd:Support for new generation of AMD SOCs.

Message ID 20210630120425.606957-1-Nehal-Bakulchandra.shah@amd.com
Headers show
Series spi:amd:Support for new generation of AMD SOCs. | expand

Message

Nehal-bakulchandra Shah June 30, 2021, 12:04 p.m. UTC 
Fix of limitation  of max 72 bytes size of fifo transfer.
Also, new generation SOC support added with modification with 
register and few of the helper functions.
As driver is only supported for X86 based platform modified the KConfig
with ACPI dependency to fix kernel test robot errors.

Changes in v3:
-Update the Kconfig with ACPI dependency

Changes in v2:
-Split the patch
-Incorporate review comments

Changes in v1:
-Initial patch for adding support of new generation of SOC -Fix for 72 bytes of FIFO Size

Nehal Bakulchandra Shah (3):
  spi:amd: Add support for latest platform
  spi:amd: Fix for transfer large size of data
  spi:amd:Fix for compilation error for non X86 platforms.

 drivers/spi/Kconfig   |   2 +-
 drivers/spi/spi-amd.c | 149 +++++++++++++++++++++++++++++++++++-------
 2 files changed, 126 insertions(+), 25 deletions(-)

Comments

Mark Brown June 30, 2021, 12:47 p.m. UTC | #1 
On Wed, Jun 30, 2021 at 05:34:25PM +0530, Nehal Bakulchandra Shah wrote:

> Update the KConfig with dependency for ACPI as driver is only
> supported for x86 platform. This fixes the compilation error
> reported by kernel test robot.

If one of the earlier patches in the series fails to build you need to
fix that patch so you don't break bisection or cause people to spend
time finding errors in the earlier patches you fix later.

In any case this doesn't do what the commit log says, ACPI is supported
on at least arm64 so if you genuinely need an x86 dependency ACPI isn't
going to cut it and COMPILE_TEST means the driver will still be built on
everything.  I don't recall anything from 0day that looked like it was
anything to do with dependencies though.
Mark Brown June 30, 2021, 3:57 p.m. UTC | #2 
On Wed, Jun 30, 2021 at 08:14:12PM +0530, Shah, Nehal-bakulchandra wrote:

> > everything.  I don't recall anything from 0day that looked like it was
> > anything to do with dependencies though.

> so now should i RESEND this patch with suggested changes,i.e removing ACPI
> depedency
> change and removing COMPILE_TEST?

No, you should fix the actual problem - like I say it looked like just a
regular coding error, not something due to an actual dependency.
Nehal-bakulchandra Shah July 27, 2021, 11:54 a.m. UTC | #3 
Hi Mark
On 6/30/2021 6:17 PM, Mark Brown wrote:
> On Wed, Jun 30, 2021 at 05:34:25PM +0530, Nehal Bakulchandra Shah wrote:

>

>> Update the KConfig with dependency for ACPI as driver is only

>> supported for x86 platform. This fixes the compilation error

>> reported by kernel test robot.

> If one of the earlier patches in the series fails to build you need to

> fix that patch so you don't break bisection or cause people to spend

> time finding errors in the earlier patches you fix later.

>

> In any case this doesn't do what the commit log says, ACPI is supported

> on at least arm64 so if you genuinely need an x86 dependency ACPI isn't

> going to cut it and COMPILE_TEST means the driver will still be built on

> everything.  I don't recall anything from 0day that looked like it was

> anything to do with dependencies though.


Thanks for reviewing this patch series, however we have decided to drop 
this patch series and our partner (Cirrus) will send the new patch 
series with

more finer changes.

Regards

Nehal Shah