arc: hsdk: bug fixes and new feature supports #19173
Conversation
bug fix in arc_connect Signed-off-by: Watson Zeng <[email protected]>
bug fix in arc_hsdk.dtsi Signed-off-by: Watson Zeng <[email protected]>
IRISZZW
requested review from
dbkinder,
ruuddw,
tbursztyka,
vanwinkeljan and
vonhust
as code owners
zephyrbot
added
area: Devicetree
area: SPI
|
All checks are passing now. Tip: The bot edits this comment instead of posting a new one, so you can check the comment's history to see earlier messages. |
|
@IRISZZW , pls the fix comment and coding style. You can refer https://docs.zephyrproject.org/latest/contribute/index.html#coding-style |
|
Hmm... I should fix the CI |
|
(FWIW, I think we should start allowing |
IRISZZW
force-pushed
the
arc_fixes
branch
2 times, most recently
from
8af74a0 to
e2ed05f
Compare
* add pinmux driver. hsdk board has arduino, mikrobus and pmod interfaces, which can be confiured for different function, such as: gpio, spi, uart, iic. * add introduction for arduino, mikrobus and pmod interfaces. Signed-off-by: Watson Zeng <[email protected]>
In some hardware, e.g. ARC HS Development kit,the peripheral space of DesignWare SPI only allowes WORD access, byte acess will raise bus error. This commit adds support for this case Signed-off-by: Watson Zeng <[email protected]>
The ARC HSDK board provides an Arduino shield interface, we can use it for TFT TOUCH SHIELD. Signed-off-by: Watson Zeng <[email protected]>
There was a problem hiding this comment.
Please care to write commit messages where you explain what was the problem you're dealing with and why did you fix it the way you did.
For example I don't understand what kind of bug "arc: hsdk: dts: bug fix" really fixes. And yes it
s not clear from the change itself. Moreover I guess HSDK's .dts was copied over from the Linux source tree and then if you really fixes a valid problem we'll need to do similar change to the same file in the Linux kernel but so far I'm not convinced there's a problem to start with.
| @@ -35,6 +35,15 @@ config SPI_DW_FIFO_DEPTH | |||
| SSI_RX_FIFO_DEPTH of the DesignWare Synchronous | |||
| Serial Interface. Depth ranges from 2-256. | |||
|
|
|||
| config SPI_DW_ACCESS_WORD_ONLY | |||
| bool "DesignWare SPI only allows word access" | |||
There was a problem hiding this comment.
I guess it has nothing to do with either DW SSI controller or with HSDK board per se but instead it depends on APB data bus configuration. So please don't mention HSDK at least here - it's a good justification for commit message but not for Kconfig - other users don't care about HSDK board at all though they might need this feature also.
Moreover in U-Boot we switched to 32-bit only accessed and so far nobody complained, see https://gitlab.denx.de/u-boot/u-boot/commit/4b5f6c52e78d43710a0d062e31de741ec76ceea1 so maybe we do the same thing here?
I.e. don't introduce yet another Kconfig option and just use 32-bit accesses all the time.
Though there seem to be some cases like picoXcell pc3x3 where 16-bit access is really required, see torvalds/linux@c4fe57f
But still if 16-bit access is not needed I'd try to go with 32-bit only for now until somebody starts to complain as this will make code cleaner and simpler for now.
This PR includes necessary bug fixes and some new feature supports of arc hsdk board. To avoid the possible confilcts of separate PR, I use one single PR to include all commits.
So please review the code in the unit of commit.