| Age | Commit message (Collapse) | Author |
|---|
|
Compared to the previous Linux 3.8 based port this adds support for
KAVERI/KABINI/MULLINS APUs and OLAND/BONAIRE/HAINAN/HAWAII GPUs.
Thanks to the OpenBSD Foundation for sponsoring this work.
|
|
attaches and claims glass console but after root is mounted if firmware
is not found efifb gets reattached to have a useable glass console.
|
|
the rule created in pf_rule_copyin(). Plugs a potential kif memory
leak in pf(4) ioctl.
OK sashan@
|
|
Constipate a couple variables. Delete IDTVEC() macro and code handling
32bit-only traps (panic instead).
ok mlarkin@
|
|
The previous version of the code used "disable_intr" as a protection
around the FPU state saving routine, which left host RFLAGS.IF=0 when
entering the guest. If the guest ended up in a non-exitable condition,
the machine could hang due to missed TLB flush IPIs (and eventual
kernel lock contention and starvation).
tested by brynet@
ok guenther@
|
|
ok mlarkin@
|
|
|
|
Found the hard way.
|
|
name coming from usbd_devinfo_vp() instead of the dummy "Generic USB
video class device" string, makes it easier to differentiate multiple
webcams in firefox's webrtc permission dialog.
ok/help mpi@
|
|
Prodded by and ok mpi@
|
|
|
|
Discussing with mpi@ and guenther@, we decided to first fix the existing
semaphore implementation with regards to SA_RESTART and POSIX compliant
returns in the case where we deal with restartable signals.
Currently we return EINTR everywhere which is mostly incorrect as the
user can not know if she needs to recall the syscall or not. Return
ECANCELED to signal that SA_RESTART was set and EINTR otherwise.
Regression tests pass and so does the posixsuite. Timespec validation
bits are needed to pass the later.
OK mpi@, guenther@
|
|
For the PRU_CONTROL bit the NET_LOCK surrounds in[6]_control() and
on the ENOTSUPP case we guard the driver if_ioctl functions.
OK mpi@
|
|
consistency.
ok guenther
note - there is no equivalent diff here for i386 since XSAVE is not used
there.
|
|
|
|
updated from userland that was marked duplicated or tentative.
Otherwise we would just lose the duplicated / tentative state and assume
that the address is now unique and usable.
OK kn
|
|
Input bluhm
OK benno, kn, claudio
|
|
|
|
malloc(9) and free(9) are MP-safe nowadays.
OK mpi@
|
|
change the memory map. diff from Ryan Lennox
ok kettenis
|
|
|
|
Ok mlarkin@ jsg@
|
|
Michael W. Bombardieri. Thanks.
|
|
Michael W. Bombardieri. Thanks.
|
|
The early freeing has become possible because aesni_process() and
crypto_freesession() are no longer serialized by the kernel lock.
The flaw has caused kernel panics with IPsec traffic.
Issue seen by some, fix tested by mabi on bugs@
OK mikeb@, mpi@
|
|
|
|
|
|
|
|
uses the O32 calling convention. Such firmware comes at least on some
Loongson 3A development boards. This differs from the LS2F-based
machines made by Lemote whose PMON uses the GCC O64 calling convention.
|
|
when indicating the boot device to the kernel. This should reduce
ambiguity in root device selection.
|
|
afterwards.
crash reported by Piotr Isajew
ok stsp@
|
|
ok guenther, jmc, tom, millert, deraadt
|
|
|
|
ok kettenis@ jmatthew@
|
|
the hardware provides crypto offload, zlib offload, and an rng.
this code only supports the rng at the moment.
this device is present on their amd seatlle platforms, and very
present on their epyc stuff.
ok kettenis@ jmatthew@
|
|
|
|
|
|
this is the same check that's done when adding normal members to
the bridge. it prevents things like gif(4) being added to the bridge,
which in turn prevents panics when shoving ethernet packets down a
layer 3 interface.
reported and tested by jasper@
|
|
|
|
cannot handle the larger firmwares, so we must place them somewhere
else where the kernel can read them. Let's use 1MB. And pay attention
if someone gets burned by this decision. Other solutions get a lot
more hairy
ok mlarkin jsing
|
|
stack buffer. With a page-aligned buffer, creating a MAP_STACK sub-region
would undo the PROT_NONE guard. Ignore that last page.
(We could check if the last page is non-RW before choosing to skip it. But
we've already elected to grow STK sizes to compensate. Always ignoring the
last page makes it a non-MAP_STACK guard page which can be opportunistically
discovered)
ok semarie stefan kettenis
|
|
pointer in the IP protocol loop. Reset the loop's mbuf when the
mbuf is freed. The latter is not necessary, but clearing a pointer
to freed data is safer.
OK dlg@
|
|
Alloc the containing struct with M_ZERO so if loadfirmware() fails and
doesn't set the pointer we won't try to free an address based on
uninitialised memory.
Use M_DEVBUF not M_DRM when freeing the buffer allocated by
loadfirmware().
|
|
operations. This resolves one of the failures frequently seen
on QEMU where spawning processes while another process is working
can lead to crashes.
Committing on behalf of drahn@
ok kettenis@ jsg@
|
|
applications will crash without this lock as other threads attempt
to walk the tree while another thread is adding/removing mappings.
Committing on behalf of drahn@
ok kettenis@
|
|
keyboard late.
This makes keyboard re-attaching due to the machdep.forceukbd sysctl(2)
respect /etc/kbdtype.
Issue reported by Artturi Alm, jmc@ and landry@.
ok kettenis@
|
|
ok visa@
|
|
ok millert@, visa@
|
|
* define V_IGN_TPR (ignore virtual TPR)
* set V_IGN_TPR while configuring window exiting (was missed in one
location along with setting of the intr vector)
This should fix unresponsive network / console issues with guests on SVM.
|
|
the brk area anyway.
- Use a larger hint bound to spread the allocations more for the 32-bit case
- Simplified the overy abstracted brs/stack allocator and switch of
guard pages for the brk case. This allows i386 some extra space,
depending on memory usage patterns.
- Reduce brk area on i386 to give the rnd space more room
ok stefan@ sthen@
|
