Re: [RFC PATCH 0/9] hw/misc: Add support for interleaved memory accesses

["Philippe Mathieu-Daudé" <f4bug@amsat.org>]

On 8/17/20 6:18 PM, Philippe Mathieu-Daudé wrote:
> This is a follow up on last year patch from Stephen Checkoway:
> "block/pflash_cfi02: Implement intereleaved flash devices"
> https://www.mail-archive.com/qemu-devel@nongnu.org/msg613077.html
> 
> Instead of making the PFLASH device more complex, I went
> implementing a generic interleaver device (that can be re-used
> for other MMIO devices, probably pointless, still useful for
> simple testing).
> 
> The series is available on [1] and is organized as:
> 
> 1/ Fix a pair of issues in memory.c as we are not exercing much
> the memory_region_[read/write]_with_attrs_accessor functions (see
> also [2]).
> 
> 2/ Allow to use qtest helpers with the 'none' machine (without any
> CPU).
> 
> 3/ Add the interleaver device
> 
> 4/ Add a device to test the interleaver device (later I'll expand
> it to other MMIO accesses, such access_with_adjusted_size, endianess
> and unaligned accesses). Let the 'none' machine create this device
> from command line.
> 
> 5/ Add qtests using 4/ to test 3/.
> 
> 6/ Add a way to display the interleaved memory (sub)regions in
> the HMP 'info mtree' command.
> 
> Based-on: <20200816173051.16274-1-f4bug@amsat.org>
> [1] https://gitlab.com/philmd/qemu/-/tree/interleaved_device-v1
> [2] https://www.mail-archive.com/qemu-devel@nongnu.org/msg730279.html

I followed Paolo's suggestion to use qtest_memread/qtest_memwrite to
avoid endianess swapping, and now cross-endianess works :)

Review still welcome, but I'll respin in September.

Regards,

Phil.