> Le 22 août 2014 à 14:29, Alexander Graf <ag...@suse.de> a écrit : > > > > > On 22.08.14 14:25, Peter Maydell wrote: > > On 22 August 2014 13:12, Alexander Graf <ag...@suse.de> wrote: > >> In Linux, the timer id is a "key" into a hash table that the kernel > >> searches to find its timer. In QEMU it's an offset into an array. > >> > >> In both cases the syscall user receives it as a token from a create > >> function and should treat it as opaque. > >> > >> So in the QEMU case it is unsigned, regardless of what the kernel allows > >> it to be, because it's an array offset. > > > > It's a number between 0 and 32. That doesn't imply that it has > > to be an unsigned variable, and we already have it in a > > signed variable arg1... > > Yes, so the end result will be the same. What's the point of this bike > shedding?
On some archs, we can imagine libc/gcc filling only the 32 lower bits (= int) of the register during the syscall, and without modifying the 32 upper bits (= garbage). You must ignore the 32 upper bits (but you can ignore the sign too). I think you can let the mask but remove the sign checking -> your patch v1 was good ... Regards, Laurent
