summaryrefslogtreecommitdiff
path: root/src/soc/amd/stoneyridge/spi.c
AgeCommit message (Collapse)Author
2018-02-16soc/amd/stoneyridge/spi: Use correct conversion specifierPaul Menzel
Use the correct conversion specifier `z` for `size_t` to fix the error below. ``` error: format '%lx' expects argument of type 'long unsigned int', but \ argument 4 has type 'size_t {aka unsigned int}' [-Werror=format=] ``` Found-by: gcc (Debian 7.3.0-3) 7.3.0 Change-Id: I05d3b6c9eec0ebf77cdb9e9928037e837f87ea03 Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-on: https://review.coreboot.org/23770 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net> Reviewed-by: Richard Spiegel <richard.spiegel@silverbackltd.com> Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2018-01-30soc/amd/stoneyridge: use new host controller programmingAaron Durbin
The SPI controller on stoneyridge apparently has a large fifo and an alternate method for programming the controller. The fifo is directly accessible as well as the rx and tx pointer in addition to the execute bit. Remove the unneeded #defines and program the host controller with the above changes in mind. In addition, add debug hooks to the driver so one can dump the state of the controller when in operation. The time it took to read 4KiB of flash in the elog driver went from 20593 microseconds to 5693 microseconds on cdx03/kahlee. BUG=b:65485690 Change-Id: Ie7ea9d18cef5511686700ad9b2b9fdfeb6d5685b Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/23493 Reviewed-by: Furquan Shaikh <furquan@google.com> Reviewed-by: Justin TerAvest <teravest@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2018-01-30soc/amd/stoneyridge: utilize full SPI flash controller fifoAaron Durbin
The spi flash host controller has a dedicated register for the opcode. Therefore, indicate to the spi subsystem that the opcode size should not be taken into account when determining maximum payload size in spi_crop_chunk(). This allows the full use of the fifo when doing transfers. BUG=b:65485690 Change-Id: Iab27a69ca72fd02bc443f0673983f3b22ffca0f5 Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/23492 Reviewed-by: Furquan Shaikh <furquan@google.com> Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Martin Roth <martinroth@google.com> Reviewed-by: Justin TerAvest <teravest@chromium.org>
2018-01-30drivers/spi: support cmd opcode deduction for spi_crop_chunk()Aaron Durbin
spi_crop_chunk() currently supports deducting the command length when determining maximum payload size in a transaction. Add support for deducting just the opcode part of the command by replacing deduct_cmd_len field to generic flags field. The two enums supported drive the logic within spi_crop_chunk(): SPI_CNTRLR_DEDUCT_CMD_LEN SPI_CNTRLR_DEDUCT_OPCODE_LEN All existing users of deduct_cmd_len were converted to using the flags field. BUG=b:65485690 Change-Id: I771fba684f0ed76ffdc8573aa10f775070edc691 Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/23491 Reviewed-by: Furquan Shaikh <furquan@google.com> Reviewed-by: Justin TerAvest <teravest@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2018-01-24soc/amd/stoneyridge/spi: do not open code existing CAR APIsAaron Durbin
The CAR APIs already exist to deal with proper type useage. Don't open code things that already exist. BUG=b:65485690 Change-Id: I09593401513f6060a30cf5c02c94d14afbe8f4fd Signed-off-by: Aaron Durbin <adurbin@chromium.org> Reviewed-on: https://review.coreboot.org/23412 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Martin Roth <martinroth@google.com> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: Justin TerAvest <teravest@chromium.org>
2017-11-21amd/stoneyridge/spi: Fix reads greater than 5 bytesMarshall Dawson
This corrects a bug in 918c8717 "amd/stoneyridge: Add SPI controller driver". Pass a pointer to din to the do_command() function so the caller's copy is correctly updated. The bug allowed reads <= 5 bytes to work correctly (3 bytes consumed in the FIFO by the address) but overwrote data in the din buffer on larger transfers. Change-Id: I32b7752f047112849871cafc9ae33c5ea1466ee1 Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com> Reviewed-on: https://review.coreboot.org/22519 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-11-17amd/stoneyridge: Add SPI controller driverMarshall Dawson
Add more definitions for the controller registers and fields. Add source that is adapted from hudson and updated for Stoney Ridge. This was tested with follow-on patches that write S3 data to flash. BUG=b:68992021 Change-Id: I61d64cfdb4fce11c068113680da7ba6a199d6893 Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com> Reviewed-on: https://review.coreboot.org/22408 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Marc Jones <marc@marcjonesconsulting.com>