On Thu, 8 Jun 2006, Ralf Baechle wrote: > Interesting that a bug of this sort manages to survive for that long. > I guess it is proof that barely anybody is using 64-bit little endian, > yet we're cursed to support it. I might conclude that barely anybody is *yet* using NPTL on 64-bit MIPS at all, for either endianness, given that most of the problems I've been finding, in glibc as well as the kernel, don't seem endian-specific and would probably show up in a glibc testsuite run for either endianness. MIPS64 NPTL is very new and seems to do a good job of showing up bugs in the three syscall interfaces. -- Joseph S. Myers joseph@xxxxxxxxxxxxxxxx
