Re: [Qemu-devel] [PATCH] hw/net/dp8393x: Avoid unintentional sign extensions on addresses

[Laurent Vivier]

Le 09/01/2017 à 19:43, Peter Maydell a écrit :
> The dp8393x has several 32-bit values which are formed by concatenating
> two 16 bit device register values. Attempting to do these inline
> with ((s->reg[HI] << 16) | s->reg[LO]) can result in an unintended
> sign extension because "x << 16" is of type 'int' even though s->reg
> is unsigned, and so if the expression is used in a context where
> it is cast to uint64_t the value is incorrectly sign-extended.
> Fix this by using accessor functions with a uint32_t return type;
> this also makes the code a bit easier to read.
> 
> This should fix Coverity issues 1307765, 1307766, 1307767, 1307768.
> 
> (To avoid having a ctda read function only used in a DPRINTF,
> we move the DPRINTF down slightly so it can use the ttda function.)
> 
> Signed-off-by: Peter Maydell <address@hidden>

Reviewed-by: Laurent Vivier <address@hidden>

FWIW, I've tested a port of this patch with Quadra 800 emulation, and it
works:

Tested-by: Laurent Vivier <address@hidden>

Laurent