Re: [PATCH 12/13] hw/pci/pci_device: Introduce ld/st_endian_pci_dma() API

[Richard Henderson <richard.henderson@linaro.org>]

On 9/30/24 00:34, Philippe Mathieu-Daudé wrote:
> Introduce the ld/st_endian_pci_dma() API, which takes an extra
> boolean argument to dispatch to ld/st_{be,le}_pci_dma() methods.
> 
> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> ---
> TODO: Update docstring regexp
> ---
>   include/hw/pci/pci_device.h | 24 +++++++++++++++++++++++-
>   1 file changed, 23 insertions(+), 1 deletion(-)
> 
> diff --git a/include/hw/pci/pci_device.h b/include/hw/pci/pci_device.h
> index ff619241a4..dc9b17dded 100644
> --- a/include/hw/pci/pci_device.h
> +++ b/include/hw/pci/pci_device.h
> @@ -300,7 +300,29 @@ static inline MemTxResult pci_dma_write(PCIDevice *dev, dma_addr_t addr,
>   
>   #define PCI_DMA_DEFINE_LDST_END(_l, _s, _bits) \
>       PCI_DMA_DEFINE_LDST(_l##_le, _s##_le, _bits) \
> -    PCI_DMA_DEFINE_LDST(_l##_be, _s##_be, _bits)
> +    PCI_DMA_DEFINE_LDST(_l##_be, _s##_be, _bits) \
> +    static inline MemTxResult ld##_l##_endian_pci_dma(bool is_big_endian, \
> +                                                      PCIDevice *dev, \
> +                                                      dma_addr_t addr, \
> +                                                      uint##_bits##_t *val, \
> +                                                      MemTxAttrs attrs) \
> +    { \
> +        AddressSpace *pci_as = pci_get_address_space(dev); \
> +        return is_big_endian \
> +               ? ld##_l##_be_dma(pci_as, addr, val, attrs) \
> +               : ld##_l##_le_dma(pci_as, addr, val, attrs); \
> +    } \

Like the address_space_* functions, I think the endianness is being handled at the wrong 
level here.  Improve sysemu/dma.h first.


r~