Re: [PATCH v4 2/4] spi: spi-altera-dfl: support n5010 feature revision

From: Martin Hundebøll
Date: Wed Jul 14 2021 - 10:27:09 EST




On 14/07/2021 16.22, Tom Rix wrote:

On 7/14/21 4:33 AM, Martin Hundebøll wrote:


On 06/07/2021 16.56, Tom Rix wrote:
diff --git a/drivers/spi/spi-altera-dfl.c b/drivers/spi/spi-altera-dfl.c
index 3e32e4fe5895..f6cf7c8d9dac 100644
--- a/drivers/spi/spi-altera-dfl.c
+++ b/drivers/spi/spi-altera-dfl.c
@@ -111,6 +111,13 @@ static struct spi_board_info m10_bmc_info = {
      .chip_select = 0,
  };
+static struct spi_board_info m10_n5010_bmc_info = {
+    .modalias = "m10-n5010",
+    .max_speed_hz = 12500000,
+    .bus_num = 0,
+    .chip_select = 0,
+};

Other then the modalias, this is exactly the same as m10_bmc_info.

Why not set platform_data?

So like this?

+static struct spi_board_info m10_n5010_bmc_info = {
+    .platform_data = "m10-n5010",
+    .max_speed_hz = 12500000,
+    .bus_num = 0,
+    .chip_select = 0,
+};

I don't see how that should improve the situation, but we might allocate
the board info on the stack and set modalias dynamically instead?

No, I mean that instead of have two *bmc_info's generalize the existing one.

This could be done by using the as yet unused platform_data field to hold the identity as a bit/enum in an int.

But the existing one is global/static, so it would need
to be moved to the function/stack, yes?


In which case we could as easily keep the existing approach
of modalias, and avoid changing both drivers/fpga/dfl-n3000-nios.c
and drivers/mfd/intel-m10-bmc.c ...


// Martin