On 11/3/25 08:31, Bernhard Beschow wrote:
Am 10. März 2025 00:06:20 UTC schrieb "Philippe Mathieu-Daudé"
<phi...@linaro.org>:
The previous commit removed the single use of instance
setting the "endianness" property.
Since classes can register their io_ops with correct
endianness, no need to support different ones.
Remove the code related to SDHCIState::endianess field.
Remove the now unused SDHCIState::io_ops field, since we
directly use the class one.
Suggested-by: Bernhard Beschow <shen...@gmail.com>
Signed-off-by: Philippe Mathieu-Daudé <phi...@linaro.org>
---
hw/sd/sdhci-internal.h | 1 -
include/hw/sd/sdhci.h | 2 --
hw/sd/sdhci.c | 33 +++------------------------------
3 files changed, 3 insertions(+), 33 deletions(-)
I don't have the code in front of me right now. IIRC, the PCI subclass sets the
"endianness" property as well, but doesn't need to. This has to be removed,
otherwise creation of the PCI device will fail.
There was a patch in v4 doing that, but I dropped it in v5 :)
