On 10/8/18 8:38 AM, Johan Hovold wrote: > On Sun, Oct 07, 2018 at 09:06:19PM -0700, H. Peter Anvin wrote: >> From: "H. Peter Anvin (Intel)" <hpa@xxxxxxxxx> >> >> Alpha has had c_ispeed and c_ospeed, but still set speeds in c_cflags >> using arbitrary flags. Because BOTHER is not defined, the general >> Linux code doesn't allow setting arbitrary baud rates, and because >> CBAUDEX == 0, we can have an array overrun of the baud_rate[] table in >> drivers/tty/tty_baudrate.c if (c_cflags & CBAUD) == 037. >> >> Resolve both problems by #defining BOTHER to 037 on Alpha. >> >> However, userspace still needs to know if setting BOTHER is actually >> safe given legacy kernels (does anyone actually care about that on >> Alpha anymore?), so enable the TCGETS2/TCSETS*2 ioctls on Alpha, even >> though they use the same structure. Define struct termios2 just for >> compatibility; it is the exact same structure as struct termios. In a >> future patchset, this will be cleaned up so the uapi headers are >> usable from libc. > > Is this really needed? By defining BOTHER (and IBSHIFT which you forgot > to mention here) you are enabling arbitrary rates also through TCSETS on > alpha, right? > Yes, it's needed, not because the old ioctls won't work on NEW kernels, but because Alpha is so far behind the times, *and* the OLD kernels are severely broken if we pass BOTHER to them, we need a new ioctl number so we can guarantee that we won't do anything that user space doesn't intend; this is actually made far worse because if I read the code correctly, the kernel will still report back BOTHER and the speed field set on a legacy kernel in response to TCGETS, but the values will be completely bogus. This means that glibc will need a workaround for Alpha only, and the new ioctl numbers handles support for it. gcc should be able to fold the code together, since it should be able to detect that multiple branches of execution are otherwise identical. To micro-optimize, we could define TERMIOS_OLD as (CBAUDEX ? 8 : 0) in a future (non-stable) patch. We don't need to worry about it on PowerPC because PowerPC implemented this so long ago, before the current glibc support threshold. -hpa
