Hi Tom, On Wed, Oct 21, 2026 at 5:00 PM Tom Rini <[email protected]> wrote: > > On Wed, Oct 21, 2026 at 04:14:42PM -0500, Anirudh Srinivasan wrote: > > Hello Tom, > > > > On Wed, Oct 21, 2026 at 3:57 PM Tom Rini <[email protected]> wrote: > > > > > > On Mon, Aug 03, 2026 at 02:00:35PM -0500, Anirudh Srinivasan via U-Boot > > > wrote: > > > > > > > The designware spi driver doesn't support dual/quad/octal modes, but > > > > the driver still tries using these modes when it sees a capable flash > > > > chip in the devicetree. Add a supports_op function to this driver to > > > > prevent this from happening and force single lane operation. > > > > > > > > Signed-off-by: Anirudh Srinivasan <[email protected]> > > > > > > This introduces a failure to build: > > > +(socfpga_secu1) arm-linux-gnueabi-ld: drivers/spi/designware_spi.o: in > > > function `dw_spi_supports_op': > > > +(socfpga_secu1) > > > drivers/spi/designware_spi.c:707:(.text.dw_spi_supports_op+0x18): > > > undefined reference to `spi_mem_default_supports_op' > > > +(socfpga_secu1) make[1]: *** [Makefile:2131: u-boot] Error 1 > > > +(socfpga_secu1) make: *** [Makefile:189: __sub-make] Error 2 > > > > Thanks for catching this. I think the solution would be for > > DESIGNWARE_SPI to select SPI_MEM. Other spi drivers that call this > > function also do this (nxp_xspi, spi-aspeed-smc, cadence_qspi....). > > > > diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig > > index cfbedd64c4c..a5623a2cec6 100644 > > --- a/drivers/spi/Kconfig > > +++ b/drivers/spi/Kconfig > > @@ -222,6 +222,7 @@ config DAVINCI_SPI > > > > config DESIGNWARE_SPI > > bool "Designware SPI driver" > > + select SPI_MEM > > help > > Enable the Designware SPI driver. This driver can be used to > > access the SPI NOR flash on platforms embedding this Designware > > > > > > I will send a v2 with this fixed. > > Yes, and please have that in its own patch as a bugfix. Thanks!
Is a bugfix required here? The designware spi driver currently doesn't rely on any of the symbols provided by CONFIG_SPI_MEM. So it doesn't need to select SPI_MEM currently, right? It's needed only because my patch introduces a dependency on that symbol (so it should be selected in my patch only). If there is something else I'm missing, please let me know.
