the value fusubail in the pcb_onfault field of the wrong process if a
profiling timer interrupt happened to occur inside cpu_switch() at a
point where curproc had been updated to point to the new process but
curpcb still pointed to the old one. trap() would then fail to ignore
any page fault in fuswintr/suswintr and the system would lock up.
This bug only affected systems with kern_clock.c 1.72 or newer, as
earlier versions only call addupc_intr from interrupts that occur in
user mode. Fixed by assigning to curproc->p_addr->u_pcb.pcb_onfault
instead of curpcb->pcb_onfault, as the former is what trap() checks.
to be very unhappy with life. Audited this file again, which should
bring it back in line, and fix the problem noted by Bill Studenmund.
Thanks for taking the time to find the bug Bill.
Set this to dv_xname for scsibus and atapibus.
Set the name of the kernel thread to chan_name instead of controller's
name:channel number (so that we can use this name for controller-specific
threads).
These targets are identical to the non-nbsd versions except that they
mark resulting ELF objects and executables as OSABI_NETBSD version 0.
* Add big-endian ARM BFD vectors to arm-*-netbsdelf. Make the -nbsd
BFD target the default for arm-*-netbsdelf.
(${KERNEL_SUFFIXES}), put the list in a shell variable and use that.
Both bash (before 2.05a) and Solaris /usr/xpg4/bin/sh report a syntax
error otherwise.
the QNX licence seems to be allow both non-commercial and commercial
use actually.
According to Darren, the H.323 proxy code is buggy ATM, but is imported
here for reference anyway.
npxprobe1(), set npxdna_func to new npxdna_empty(). This function just
returns 0 and thus fallbacks to a T_DNA trap (and math emulate, if present).
This fixes kern/15828 by Masaru Oki, tested on 386DX machine.
Reviewed by Jason Thorpe.
make npxdna_s87() and npxdna_xmm() static
g/c (npx_type == NPX_NONE) test from npxdna_s87(), it's no longer needed
- If VOP_ACCESS fails when updating mount, we will vrele() twice.
- The check for update-only flags in mp->mnt_flag when not updating
case is bogus. If we really want to check, we need to see flags in
ufs_args, but I'm not sure if it is really necessary.
- The credential passed to ffs_reload was credential of when looking
up mount point, but now it is credential of when looking up device
node. Anyway, it may be current process's credential.
briggs