On 6/3/25 19:23, BALATON Zoltan wrote:
On Mon, 3 Mar 2025, BALATON Zoltan wrote:
On Mon, 3 Mar 2025, Philippe Mathieu-Daudé wrote:
Hi Zoltan,
On 10/2/25 17:03, BALATON Zoltan wrote:
The interrupt enable registers are not reset to 0 on Freescale eSDHC
but some bits are enabled on reset. At least some U-Boot versions seem
to expect this and not initialise these registers before expecting
interrupts. Use existing vendor property for Freescale eSDHC and set
the reset value of the interrupt registers to match Freescale
documentation.
Signed-off-by: BALATON Zoltan <bala...@eik.bme.hu>
---
v2: Restrict to e500. Adding a reset method in a subclass does not
work because the common reset function is called directly on register
write from the guest but there's already provision for vendor specific
behaviour which can be used to restrict this to Freescale SoCs.
hw/ppc/e500.c | 1 +
hw/sd/sdhci.c | 4 ++++
include/hw/sd/sdhci.h | 1 +
3 files changed, 6 insertions(+)
diff --git a/hw/ppc/e500.c b/hw/ppc/e500.c
index 26933e0457..560eb42a12 100644
--- a/hw/ppc/e500.c
+++ b/hw/ppc/e500.c
@@ -1044,6 +1044,7 @@ void ppce500_init(MachineState *machine)
dev = qdev_new(TYPE_SYSBUS_SDHCI);
qdev_prop_set_uint8(dev, "sd-spec-version", 2);
qdev_prop_set_uint8(dev, "endianness", DEVICE_BIG_ENDIAN);
+ qdev_prop_set_uint8(dev, "vendor", SDHCI_VENDOR_FSL);
s = SYS_BUS_DEVICE(dev);
sysbus_realize_and_unref(s, &error_fatal);
sysbus_connect_irq(s, 0, qdev_get_gpio_in(mpicdev,
MPC85XX_ESDHC_IRQ));
diff --git a/hw/sd/sdhci.c b/hw/sd/sdhci.c
index 99dd4a4e95..afa3c6d448 100644
--- a/hw/sd/sdhci.c
+++ b/hw/sd/sdhci.c
@@ -307,6 +307,10 @@ static void sdhci_reset(SDHCIState *s)
s->data_count = 0;
s->stopped_state = sdhc_not_stopped;
s->pending_insert_state = false;
+ if (s->vendor == SDHCI_VENDOR_FSL) {
+ s->norintstsen = 0x013f;
+ s->errintstsen = 0x117f;
I'd rather do like capareg, and add:
DEFINE_PROP_UINT16("norintstsen", _state, norintstsen, 0),
...
I don't see what you mean. capareg does not seem to be set via a
property.
Then SoC code sets it:
qdev_prop_set_uint16(dev, "norintstsen", 0x013f);
...
WDYT?
I think it may be overkill to add properties for this if there are no
other vendor or variant that needs this. Also properties are for
something the user may want to set as those can be changed with QEMU
command line and these reset values aren't something the user should
change so I think this patch is the simplest solution now.
This patch wasn't in the pull request but I haven't seen an answer to
this message either so was it missed or do you have furhter comments?
Bernhard has a comment about naming of SDHCI_VENDOR_FSL but I think the
already existing IMX name is what's wrong not the one added in this
patch but I don't think that's really that confusing to worth further
effort. We still have time as this can be considered a fix but I'd like
this to not get forgotten so I bring it up again.
I'll post a v3 with what I had in mind.
