| Age | Commit message (Collapse) | Author |
|---|
|
This adds stacktrace_save_utrace() to extract and save the userland stack
which is stubbed out on most archs. alpha and riscv64 do not even implement
dt(4) and stacktrace_save_at() so the stubs are excluded there.
Additionally add a new ioctl DTIOCGETAUXBASE which allows btrace to
fetch the AUX_BASE vallue from the AUX vector of a process.
OK mpi@ (some time ago) discussed with kettenis@
|
|
|
|
Obsolete since last year's r1.411 "Rework the rttimer code."
OK claudio
|
|
----------------------------
revision 1.341
date: 2023/04/26 09:39:56; author: dv; state: Exp; lines: +3 -3; commitid: 1XztmVbxa8OP2xYi;
Unbreak vmm on Intel. Mistakenly enabled CET on non-CET systems.
Found by anton@ & jsg@
----------------------------
revision 1.340
date: 2023/04/24 16:53:57; author: dv; state: Exp; lines: +11 -1; commitid: q9qQ55f7h2gen8N4;
vmm(4): allow guests to enable and use supervisor IBT.
Why should hosts have all the fun? Conditionally unmask the cpuid
bits for IBT and allow r/w access to the supervisor CET msr.
Will need revisiting when we introduce usage of userland CET msr.
|
|
|
|
---
revision 1.339
date: 2023/04/22 18:27:28; author: guenther; state: Exp; lines: +3 -3; commitid: ajtkYPSAhtJdB488;
Rename the XCR0_* #defines to XFEATURE_* and add the new supervisor-state
features: while all are appropriate for xsaves/xrstors, the
supervisor-state features aren't for xcr0 but rather for the new XSS_MSR,
making the current names kinda confusing.
Add #defines for masking bits for xcr0 vs XSS.
Add and report the new XSAVE_XFD xsave subfeature bit.
---
|
|
|
|
Called wrong incore when not holding bio.
ok claudio@
|
|
ok dv, deraadt
|
|
|
|
ok deraadt@, patrick@
|
|
Do not grab the `vmmaplk' recursively, prevent a self-deadlock.
It causes panic: uvm_map_pageable_wire: stale map
Found by regress/misc/posixtestsuite conformance/interfaces/mmap/18-1
requested by deraadt@
|
|
|
|
noticed by miod@
ok kettenis@
|
|
ok claudio@ kettenis@
|
|
Found by anton@ & jsg@
|
|
ok jsg@
|
|
|
|
This simplifies syzkaller revival after the removal of __syscall.
OK bluhm, millert, deraadt
|
|
with tweaks from bluhm, claudio and dlg
I fine with it from claudio
looks good to me from dlg
ok bluhm
|
|
|
|
|
|
|
|
P_HASSIBLING() using this count.
OK mvs@ mpi@
|
|
|
|
So kernel lock is only needed for changing the route rt_flags. In
arpresolve() protect rt_llinfo lookup and llinfo_arp modification
with arp_mtx. Grab kernel lock for rt_flags reject modification
only when needed.
Tested by Hrvoje Popovski; OK patrick@ kn@
|
|
doesn't know where to send traffic. This will send an unsolicited
neighbor advertisement, as described in RFC9131, to the all-routers
multicast address so all routers on the same link will learn the path
back to the address.
This is intended to speed up the first return packet on an IPv6 interface.
OK florian@
|
|
Bootloaders have no write support for softraid volumes, which manifests in,
e.g. /bsd.upgrade not being stripped of its 'x' permission bit to prevent
sysupgrade loops in case of upgrade failure.
Set a no-write flag handled by libsa to bail out early in write calls.
There should be no real behaviour change, writes just fail earlier now.
i386 BIOS. Crank minor.
Tested inside amd64 vmm.
|
|
GENERIC, efiboot and installboot(8) all have softraid enabled, the manual
already documents boot support for riscv64, "just" bsd.rd lacks it.
(thought I had committed this months ago...)
SiFive HiFive Unmatched A00.
OK patrick
|
|
Bootloaders have no write support for softraid volumes, which manifests in,
e.g. /bsd.upgrade not being stripped of its 'x' permission bit to prevent
sysupgrade loops in case of upgrade failure.
Set a no-write flag handled by libsa to bail out early in write calls.
There should be no real behaviour change, writes just fail earlier now.
sparc64 ofwboot. Crank minor.
Tested on T4-2 guest domain with root on softraid.
|
|
The object sent to vmm(4) contained file paths and details the
kernel does not need for cpu virtualization as device emulation is
in userland. Effectively, "pull up" the struct members from the
vm_create_params struct to the parent vmop_create_params struct.
This allows us to clean up some of vmd(8) and simplify things for
switching to having vmctl(8) open the "kernel" file (SeaBIOS, bsd.rd,
etc.) to allow users to boot recovery ramdisk kernels.
ok mlarkin@
|
|
Change the semantic of vm_map_busy() to be able to completely unlock the
`vmmaplk' instead of downgrading it to a read lock in mlock(2). This is
necessary because uvm_fault_wire() tries to re-grab the same lock.
We now keep track of the thread currently holding the vmmap busy to ensure
it can relock & unbusy the vmmap. The new pattern becomes:
....vm_map_lock(map);
....vm_map_busy(map); /* prevent other threads to grab an exclusive lock */
....vm_map_unlock(map);
....
..../*
.... * Do some stuff generally requiring a tsleep(9).
.... */
....
....vm_map_lock(map);
....vm_map_unbusy(map); /* allow other threads to make progress after unlock */
....vm_map_unlock(map);
Fix adapted from NetBSD's r1.249 of uvm/uvm_map.c. Issue reported by
Jacqueline Jolicoeur exposed by a "wallet refresh" of the Monero App.
Panic hand-copied below:
sleep_finish()
rw_enter()
uvmfault_lookup()
uvm_fault_check()
uvm_fault()
uvm_fault_wire()
uvm_map_pageable_wire()
sys_mlock()
ok kettenis@
|
|
makefiles.
Fixes & ok kn@
No intentional functional change.
|
|
ok kettenis@
|
|
the display turns off and it actually looks like it's properly suspended.
ok kettenis@ tobhe@
|
|
restore on wakeup.
ok patrick@ kettenis@
|
|
to the LPIs. So far we have only flipped the enable bit in the shared
memory, but it turns out that the state is actually cached in the ITS.
We have to send an invalidation to flush it, both after disabling and
enabling the LPI. We probably also should do it after establishing an
interrupt, and especially when disestablishing, e.g. for PCIe hotplug.
Since we do not currently store the device/event IDs we cannot do a
targeted invalidation and have to fall back to sending one that clears
the whole cache to each ITS. In the future we should keep some more
state per established LPI.
Furthermore we currently keep a list of LPIs in each ITS instance, which
means in case we have multiple ITS we will re-use LPI numbers. But since
there's no relevant HW so far that has those, there's no rish to fix that.
This fixes suspend/resume on x13s with NVMe+MSI.
ok kettenis@
|
|
Bootloaders have no write support for softraid volumes, which manifests in,
e.g. /bsd.upgrade not being stripped of its 'x' permission bit to prevent
sysupgrade loops in case of upgrade failure.
Set a no-write flag handled by libsa to bail out early in write calls.
There should be no real behaviour change, writes just fail earlier now.
amd64 EFI bits, other architectues to come. Crank minor.
"seems sensible" jsing
|
|
Bootloaders have no write support for softraid volumes, which manifests in,
e.g. /bsd.upgrade not being stripped of its 'x' permission bit to prevent
sysupgrade loops in case of upgrade failure.
Set a no-write flag handled by libsa to bail out early in write calls.
There should be no real behaviour change, writes just fail earlier now.
amd64 BIOS bits first, rest to come. Crank minor.
"seems sensible" jsing
|
|
|
|
and Samsung PM9B1 as found in sthen's t14 gen 3 intel
|
|
On MULTIPROCESSOR systems, the following sequence of kgmon(8)
invocations leaves the statclock() frequency at stathz when there is
still a CPU on the system where the gmon state is GMON_PROF_ON:
# kgmon -c 0 -b
# kgmon -c 1 -b
# kgmon -c 0 -h
The problem is that we aren't counting CPUs with profiling enabled.
Add "gmon_cpu_count" to keep a count. Call startprofclock() for the
first CPU to enable profiling and stopprofclock() for the last CPU to
disable profiling.
|
|
Add a third parameter to addupc_intr(), "u_long nticks". This will
allow us to credit more than one profiling tick to the thread at once.
Should be useful in the unusual case where the clock interrupt is
masked for an extended period.
|
|
Why should hosts have all the fun? Conditionally unmask the cpuid
bits for IBT and allow r/w access to the supervisor CET msr.
Will need revisiting when we introduce usage of userland CET msr.
ok marlkin@
|
|
in preparation for improvements.
ok claudio@
|
|
This will make MSIs work on the Lenovo x13s as soon as an updated device tree
is installed.
Discussed with kettenis@
|
|
binding in Linux 6.3 device trees. While there, remove retrieving the register
address, as the new binding has an additional register with doesn't work with
a single OF_getpropint(), and since we're not using it we don't need to keep it
around.
|
|
in6.c already has the privilege check as early as possible, make in.c match.
For unprivileged IPv4 ioctl calls with invalid args, this changes errno from
E* to EPERM.
OK bluhm
|
|
ok robert@, deraadt@
|
|
or not. The idea is that since /usr/local has wxallowed by default this
will enable enforcement for base while leaving ports alone for now. This
will help us transition to a state where ports are properly marked and
allow us to establish that base is really clean.
Also add an exception for chrome. Chrome already appears to be clean on
arm64 and this exception can be easily modified for testing other ports.
This will screw over people that deliberately disable wxallowed on
/usr/local or who don't have a separate partition for /usr/local. We
think that is an acceptable compromise for the next months.
ok robert@, deraadt@ (who came up with the idea)
|
