This patch reverts the changes made in D19670 and fixes the original
issue by allocating and prepopulating a leaf page table page for wired
userspace 2M pages.
The original issue is an edge case that creates an unmapped, wired
region in userspace. Subsequent faults on this region can trigger wired
superpage creation, which leads to a panic in pmap_demote_pde_locked()
as the pmap does not create a leaf page table page for the wired
superpage. D19670 fixed this by disallowing preemptive creation of
wired superpage mappings, but that fix is currently interfering with an
ongoing effort of speeding up vm_map_wire for large, contiguous entries
(e.g. bhyve wiring guest memory).
Reviewed by: alc, markj
Sponsored by: Google, Inc. (GSoC 2023)
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D41132
They are a bit more informative than raw hexadecimal values.
While here, sort existing defines of bits for AMD MSRs to match the address
order.
Reviewed by: kib, emaste
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D41816
To help porting the Linux emulation layer to a new platforms start using
Linux names for conditional builds instead of architecture-specific ifdefs.
MFC after: 1 week
Switch to using db_addr_t to hold frame pointer values until they are
verified to be suitably aligned.
Reviewed by: kib, markj
Differential Revision: https://reviews.freebsd.org/D41532
For amd64, i386, arm, and riscv, i.e. all architectures except arm64,
the custom implementation is provided since we maintain the bitmask of
active CPUs anyway.
Arm64 uses somewhat naive iteration over CPUs and match current vmspace'
pmap with the argument. It is not guaranteed that vmspace->pmap is the
same as the active pmap, but the inaccuracy should be toleratable.
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D32360
Adding a writev syscall wrapper is needed due to Linux family of write
syscalls doesn't distinguish between in kernel blocking operations
and always returns EAGAIN while FreeBSD can return ENOBUFS.
MFC after: 1 month
Adding a writev syscall wrapper is needed due to Linux family of write
syscalls doesn't distinguish between in kernel blocking operations
and always returns EAGAIN while FreeBSD can return ENOBUFS.
MFC after: 1 month
Adding a write syscall wrapper is needed due to Linux family of write
syscalls doesn't distinguish between in kernel blocking operations
and always returns EAGAIN while FreeBSD can return ENOBUFS.
MFC after: 1 month
Adding a write syscall wrapper is needed due to Linux family of write
syscalls doesn't distinguish between in kernel blocking operations
and always returns EAGAIN while FreeBSD can return ENOBUFS.
MFC after: 1 month
bhyve in a 13.x jail fails to boot guests with more than one vCPU
because they pass too small a buffer to VM_GET_CPUS, causing the ioctl
handler to return ERANGE. Handle this the same way as cpuset system
calls: make sure that the result can fit in the truncated space, and
relax the check on the cpuset buffer.
As a side effect, fix an insufficient bounds check on "size". The
signed/unsigned comparison with sizeof(cpuset_t) fails to exclude
negative values, so we can end up allocating impossibly large amounts of
memory.
Reviewed by: jhb
MFC after: 1 week
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D41496
EARLY_AP_STARTUP was introduced in 2016 (commit fdce57a042) with note:
As a transition aid, the new behavior is moved under a new
kernel option (EARLY_AP_STARTUP). This will allow the option
to be turned off if need be during initial testing. I hope to
enable this on x86 by default in a followup commit ...
It was enabled by default, but became effectively mandatory (on x86)
some time later. Move it to DEFAULTS to avoid an unbootable system if
the option is left out of a custom kernel configuration file.
Reported by: wollman
Reviewed by: jhb
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D41352
Following the removal of general MIPS support, there's no longer a need
to have the AHB bus-frontend in place, which according to Linux sources
also isn't used with any non-MIPS SoCs. For simplicity, PCI bus support
is only made conditional on the main one again, i. e. device ath_pci is
removed, and built into the main module, i. e. if_ath_pci.ko obsoleted,
respectively.
Effectively, this reverts the following commits and associated changes:
dba9c85977e849bb3ecb
Approved by: adrian
Relnotes: yes
Differential Revision: https://reviews.freebsd.org/D41354
Because KASAN shadows the kernel image itself (KMSAN currently does
not), a shadow mapping of the boot stack must be created very early
during boot. pmap_san_enter() reserves a fixed number of pages for the
purpose of creating and mapping this shadow region.
After commit 789df254cc ("amd64: Use a larger boot stack"), it could
happen that this reservation is insufficient; this happens when
bootstack crosses a PAGE_SHIFT + KASAN_SHADOW_SCALE_SHIFT boundary.
Update the calculation to take into account the new size of the boot
stack.
Fixes: 789df254cc ("amd64: Use a larger boot stack")
Sponsored by: The FreeBSD Foundation
Hardware with more than 256 CPU cores is currently available and will
become increasingly common over FreeBSD 14's lifetime. Increase MAXCPU
in the amd64 GENERIC kernel configuration to 1024.
Earlier commits increased some related limits. These prerequisite
commits include at least:
- d7ed40243769 Increase MAX_APIC_ID safeguard to 0x800
- d1639e43c5 cpuset: increase userland maximum size to 1024
Global and allocated arrays sized by MAXCPU result in excessive bloat
on systems with lower core counts. In addition, some code used u_char
(8 bits) to hold a CPU index, which is not valid if MAXCPU is greater
than 256.
A number of recent commits addressed these sorts of issues, including
at least:
- 133935d26f pf: atomically increment state ids
- 74ac712f72 vmm: Dynamically allocate a couple of per-CPU state save areas
- 78cfa762eb callout: Move per-CPU callout state into the dpcpu region
- 42f722e721 amd64: store pcids pmap data in pcpu zone
- 9801e7c275 smp_topo: dynamically allocate group array
- 9fb6718d1b smp: Dynamically allocate the stoppcbs array
- 2bb16c6352 x86: retire use of intr_bind
There are some additional allocations still to be converted and
more scalability work is required to make effective use of very high
core count systems, but this change allows us to boot on these systems
and provides a Kernel Binary Interface (KBI) for the FreeBSD 14 release
that supports these configurations.
Special thanks to AMD for providing hardware to test these changes.
PR: 269572
Reviewed by: des
Relnotes: Yes
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D36838
amd64 is special in that its implementation of zpcpu_offset_cpu() is not
the identity transformation, even in !SMP kernels. Because the pm_pcidp
array of amd64's struct pmap is allocated from a pcpu UMA zone, this
means that accessing pm_pcidp directly, as is done in !SMP
implementations of pmap_invalidate_*, does not work. Specifically, I
see occasional unexplicable crashes in userspace when PCIDs are enabled.
Apply a minimal patch to fix the problem. While it would also make
sense to provide separate implementations of zpcpu_* for !SMP kernels,
fixing it this way makes the SMP and !SMP implementations of
pmap_invalidate_* more similar.
Reviewed by: alc, kib
MFC after: 1 week
Sponsored by: Klara, Inc.
Sponsored by: Juniper Networks, Inc.
Differential Revision: https://reviews.freebsd.org/D41230
On x86 Linux via AT_HWCAP2 the user controlled (by tunables) processor
capabilities are exposed.
Reviewed by:
Differential Revision: https://reviews.freebsd.org/D41165
MFC after: 2 weeks
With sanitizers enabled, it becomes possible to overflow the stack when
only a single page is used. Follow arm64's example and use the default
kernel stack size instead. This is a bit wasteful, but without a guard
page, overflow merely corrupts adjacent .bss entries and is thus
difficult to debug.
Note, with a GENERIC kernel we already consume over half of the
available boot stack space, see the review for an example.
Reviewed by: kib
Reported by: Jenkins
MFC after: 2 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D41166
Several vm_radix tries are not initialized with vm_radix_init. That
works, for now, since static initialization zeroes the root field
anyway, but if initialization changes, these tries will fail. Add
missing initializer calls.
Reviewed by: alc, kib, markj
Differential Revision: https://reviews.freebsd.org/D40971
The function pmap_pde_ept_executable() should not be conditionally
compiled based on VM_NRESERVLEVEL. It is required indirectly by
pmap_enter(..., psind=1) even when reservation-based allocation is
disabled at compile time.
Reviewed by: alc
MFC after: 1 week
