s/mipsNN_cp0_cause_read/mips_cp0_cause_read/ in INGENIC_INTR_DEBUG guarded code.
It is disabled though, thus moving #ifdef 0 higher to make option build.
(otherwise buffer is unused and compiler complains)
Not sure why the code was disabled, potentially it was too verbose.
Replace MFC0 macro with CP0 accessor functions in debug code.
Include ingenic_var.h in ingenic/mainbus.c when INGENIC_DEBUG is enabled and
move "opt_ingenic.h" higher so this option is visible for the include.
MFC0 macro was removed in ingenic_regs.h v1.25, but debug code was not adopted.
Fixes INGENIC_DEBUG enabled build for the evbmips CI20 kernel config.
Rename the "basic 4BSD" disklabel routines readdisklabel_4bsd()
and writedisklabel_4bsd(), but add weak aliases to the standard
names. This easily allows machine-specific code to look for an
odd-ball machine-firmware-specific format that might be required
only for booting but otherwise use garden-variety BSD disklabels.
nvmm(4): Fix race in clearing vcpu->comm->stop bit.
Don't touch vcpu after nvmm_vcpu_put when we no longer hold the lock,
and certainly not after nvmm_machine_put when the vcpu could even be
freed.
PR kern/60544: nvmm_vcpu_run: use after mutex release
nvmm(4) vmx: Don't read VMCS_EXIT_REASON if vmlaunch/vmresume failed.
It is not valid in that case; we might end up operating on the
_previous_ successful vmlaunch/vmresume's now-stale exit reason.
kern/60545: nvmm_x86_vmx: stale exit reason on launch/resume failure
proc_regio: Clarify control flow and sprinkle comments.
Use `if (error) goto out' instead of `if (error == 0) error = ...' to
make the error case non-linear and indented, and the success case a
straight line of logic.
No functional change intended.
PR port-amd64/60556: panic in process_read_fpregs_xmm
ptrace/procfs: Take lwp lock around process_read/write_(fp)regs(32).
And verify that the lwp is in LSSTOP state before trying to read.
Otherwise, we might catch an lwp while it's still running on the CPU,
with all FPU state active, making the result unstable and tripping
assertions on some architectures.
PR port-amd64/60556: panic in process_read_fpregs_xmm
grep(1): Nix xfail on new test for PR 60552 screw case.
I was planning to do this in the same commit that fixes it, to keep
the history of bugs and fixes tidy, but gutteridge@ beat me to
committing the fix.
PR bin/60552: /usr/bin/grep failure under AMD64
nfs: Avoid UB in integer shift amounts.
Just use ilog2 explicitly here, instead of writing out a loop with
shifts and increments to compute it more slowly and buggily. Note
the confusing fencepost adjustment. Clamp _PC_FILESIZEBITS at the
actual size of off_t since we can't go beyond that.
PR kern/60554: nfs: integer shift UB in pathconf(_PC_FILESIZEBITS)
Make console input on virt68k 8-bit clean.
On signed-char ports you can't feed a "char *" to the tty layer. A cursory
glance for more instances of this didn't find any but it's possible.
Simplify the interface between the timers and the clock infrastructure,
and allow the clock infrastructure to select clock roles without having
netbsd,hardclock or netbsd,statclock properties in /chosen (they're now
optional).
Add fdtbus_get_reg_shift() to get the "reg-shift" property and provide
a default value, rather than open-coding it everywhere it's used. It's
common enough in various device bindings, and ePAPR describes it as a
generic device property.
Deal with the absolutely pathological handling of the IDENTIFY command
on big-endian systems.
In wdc_datain_pio(), in the ATA_DRIVE_NOSTREAM case, if the bus.h
implementation does does not have __BUS_SPACE_HAS_STREAM_METHODS
and the system is _BIG_ENDIAN, individually swap each 16-bit value
rather than assuming the bus_space_read_*2() implementation will do
it.
This is necessary because IDENTIFY is treated differently than every
other command; all callers assume it will be returned in host-order,
when in reality it's an array of 256 16-bit little-endian integers (even
the string fields).
Ultimately, the correct fix is to stop special-casing IDENTIFY, but
that is a much larger change due to the confusing array of flags that
control I/O size and disposition, and requires much wider testing.
