Re: [PATCH] softmmu/memory: use memcpy for multi-byte accesses

[Patrick Venture <venture@google.com>]

[-- Attachment #1: Type: text/plain, Size: 1428 bytes --]

On Wed, Nov 15, 2023 at 2:30 AM Peter Maydell <peter.maydell@linaro.org>
wrote:

> On Tue, 14 Nov 2023 at 21:18, Richard Henderson
> <richard.henderson@linaro.org> wrote:
> >
> > On 11/14/23 12:55, Patrick Venture wrote:
> > > Avoids unaligned pointer issues.
> > >
> > > Reviewed-by: Chris Rauer <crauer@google.com>
> > > Reviewed-by: Peter Foley <pefoley@google.com>
> > > Signed-off-by: Patrick Venture <venture@google.com>
> > > ---
> > >   system/memory.c | 16 ++++++++--------
> > >   1 file changed, 8 insertions(+), 8 deletions(-)
> > >
> > > diff --git a/system/memory.c b/system/memory.c
> > > index 304fa843ea..02c97d5187 100644
> > > --- a/system/memory.c
> > > +++ b/system/memory.c
> > > @@ -1343,16 +1343,16 @@ static uint64_t
> memory_region_ram_device_read(void *opaque,
> > >
> > >       switch (size) {
> > >       case 1:
> > > -        data = *(uint8_t *)(mr->ram_block->host + addr);
> > > +        memcpy(&data, mr->ram_block->host + addr, sizeof(uint8_t));
> >
> >
> > This is incorrect, especially for big-endian hosts.
> >
> > You want to use "qemu/bswap.h", ld*_he_p(), st*_he_p().
>
> More specifically, we have a ldn_he_p() and stn_he_p() that
> take the size in bytes of the data to read, so we should be
> able to replace the switch-on-size in these functions with
> a single call to the appropriate one of those.
>

Thanks!


>
> thanks
> -- PMM
>

[-- Attachment #2: Type: text/html, Size: 2404 bytes --]