|
From: Nicholas N. <nj...@ca...> - 2004-11-16 10:02:12
|
On Tue, 16 Nov 2004, Paul Mackerras wrote: > Nicholas Nethercote writes: > >> Converted more syscalls, including futex() which I'm not too sure about. > > What aren't you sure about? Whether the wrapper is doing the right thing. In particular, whether the PRE_REG_READ and PRE_MEM_READ/PRE_MEM_WRITE calls are right. It seems futex() works in a couple of different ways depending on the 2nd 'op' arg, and I think the wrapper might be too simplistic, ie. get some cases wrong. The overloading of the 4th 'utime' param (as implied by the wrapper) looks very strange. The situation is complicated by the fact that the prototype given in the man page doesn't match that present in the kernel source code. N |