On Wed, Jun 03, 2020 at 01:05:18AM -0700, Alex Nemirovsky wrote:
From: Aaron Tseng aaron.tseng@cortina-access.com
Add Cortina Access Ethernet device driver for CAxxxx SoCs. This driver supports only the DM_ETH network model.
Signed-off-by: Aaron Tseng aaron.tseng@cortina-access.com Signed-off-by: Alex Nemirovsky alex.nemirovsky@cortina-access.com
CC: Joe Hershberger joe.hershberger@ni.com CC: Abbie Chang abbie.chang@Cortina-Access.com CC: Tom Rini trini@konsulko.com
Changes in v4: None Changes in v3:
- Changed commit comment to state that only DM model is supported
- Removed blank line at end of C file
Changes in v2:
- Remove legacy mode support
- Add support for additional SoC variants
- Remove unused variables
MAINTAINERS | 4 + drivers/net/Kconfig | 7 + drivers/net/Makefile | 1 + drivers/net/cortina_ni.c | 1909 ++++++++++++++++++++++++++++++++++++++++++++++ drivers/net/cortina_ni.h | 592 ++++++++++++++ 5 files changed, 2513 insertions(+) create mode 100644 drivers/net/cortina_ni.c create mode 100644 drivers/net/cortina_ni.h
So, is there a similar driver in upstream Linux? This driver doesn't quite fit right. At an "easy" level, there's still the customer macro around debug statements and not using pr_debug(), and contains a whole lot of debug code. There's also code for a saturn platform, is that being upstreamed? There's a ton of union usage which is also uncommon. A small item is it looks like this has its own crc function, when we have those available already. There's also things like:
+enum ca_status_t ca_mdio_read(CA_IN unsigned int addr,
CA_IN unsigned int offset,CA_OUT unsigned short *data)+{
Where CA_IN / CA_OUT just don't fit.
Which brings me back to why I asked the first question, this feels a lot like a driver for an RTOS or some other system was adapted to U-Boot, rather than writing a new driver for U-Boot based on internal knowledge of the part in question. And I think that applies to a lot of the drivers as seen in the NAND review as well. Thanks!