On 12/8/25 15:20, Alex Bennée wrote:
Philippe Mathieu-Daudé <phi...@linaro.org> writes:
Commit 1585ab9f1ba ("hw/sd/sdcard: Fill SPI response bits in card
code") exposed a bug in the SPI adapter: if no SD card is plugged,
we shouldn't return any particular packet response, but the noise
shifted on the MISO line. Return the dummy byte, otherwise we get:
qemu-system-riscv64: ../hw/sd/ssi-sd.c:160: ssi_sd_transfer: Assertion
`s->arglen > 0' failed.
Reported-by: Guenter Roeck <li...@roeck-us.net>
Fixes: 775616c3ae8 ("Partial SD card SPI mode support")
Signed-off-by: Philippe Mathieu-Daudé <phi...@linaro.org>
---
hw/sd/ssi-sd.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/hw/sd/ssi-sd.c b/hw/sd/ssi-sd.c
index 594dead19ee..3aacbd03871 100644
--- a/hw/sd/ssi-sd.c
+++ b/hw/sd/ssi-sd.c
@@ -89,6 +89,10 @@ static uint32_t ssi_sd_transfer(SSIPeripheral *dev, uint32_t
val)
SDRequest request;
uint8_t longresp[5];
+ if (!sdbus_get_inserted(&s->sdbus)) {
+ return SSI_DUMMY;
+ }
+
Seems fair although it's hard to track what is consuming this value. I
think we end up in ssi_transfer() which a surprising number of calls
don't even bother checking the return value, other just seem to | the
result when iterating across devices.
A SPI transaction consists of shifting bit in sync the CLK line,
writing on the MOSI (output) line / and reading MISO (input) line.
IOW, each time you write a word, you also read it at the same time.
When a driver just wants to write, it is OK to ignore the returned
values.
In this case, we don't want to return "There is a card with error"
because there is no card. We shift the request on the MOSI line,
but nothing replies on the MISO line.
Should I reword the commit description?
We should probably improve on the definitions of transfer/transfer_raw
and explain what the return value is.
Agreed.
Anyway:
Reviewed-by: Alex Bennée <alex.ben...@linaro.org>
Thanks!
