Andrew Turner [Thu, 27 Oct 2022 16:21:02 +0000 (17:21 +0100)]
Only build if_mvneta.c when FDT is defined
This is only ever used by a FDT specific attachment and calls into
FDT specific functions. Only build it when FDT is in the kernel config.
Sponsored by: The FreeBSD Foundation
Andrew Turner [Tue, 18 Oct 2022 16:37:43 +0000 (17:37 +0100)]
Remove an unneeded cpu_dcache_wb_range from arm64
The cpu_dcache_wb_range function is an expensive function that is
unneeded in ddb. It is used when the cache needs to be written to RAM,
e.g. when working with a non-cache coherent device.
Remove it as cpu_icache_sync_range already has the needed d-cache
handling to ensure any changed memory is visible to the i-cache.
Reviewed by: imp
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D37037
Ed Maste [Thu, 27 Oct 2022 14:49:51 +0000 (10:49 -0400)]
netlink: use (void) for function definitions with no arguments
For some of these Clang produced a warning that "a function declaration
without a prototype is deprecated in all versions of C". In other cases
the function defintion used () which did not match the header
declaration, which used (void).
arm64: Hyper-V: enablement for ARM64 in Hyper-V (Part 3, final)
This is the last part for ARM64 Hyper-V enablement. This includes
commone files and make file changes to enable the ARM64 FreeBSD
guest on Hyper-V. With this patch, it should be able to build
the ARM64 image and install it on Hyper-V.
Kyle Evans [Thu, 27 Oct 2022 03:46:45 +0000 (22:46 -0500)]
psci: set psci_present as early as possible
psci_attach is way too late to provide the intended semantics for
psci_present. psci calls can be made immediately after psci_init(),
called way earlier at SI_SUB_CPU + SI_ORDER_FIRST, and we need it to
be valid as early as we can possibly call a psci function.
This fixes booting RPi3+4 with the in-review spintable patch;
rpi3-psci-monitor patches the FDT to add a PSCI node, but it doesn't
patch each cpus' enable-method. Because of this, we would stall the
boot while enabling CPU 1 as we saw a valid looking enable-method and
"no" functional PSCI and attempted to use the spintable rather than
simply not starting secondary APs.
Refactor the symlink and mountpoint traversal logic to avoid
repeatedly checking the vnode type; a symlink cannot be a mountpoint
and vice versa. Avoid repeatedly checking cn_flags for NOCROSSMOUNT
and simplify the check which determines whether the vnode is a
mountpoint.
Remove witness directives from crossmp locking VOPs
These are of limited use since the crossmp vnode locking ops have not
actually used a lock since commit a2d35545429117e68fbcbc68e14ad55e84265d69. We in fact require that
these operations are always issued with LK_SHARED. Additionally,
these directives can produce a false positive in certain VV_CROSSLOCK
cases which require upgrading of the covered vnode lock from shared
to exclusive.
While here, replace the runtime check of LK_SHARED with a KASSERT and
expand the check to include LK_NOWAIT, which all callers pass.
Add VV_CROSSLOCK vnode flag to avoid cross-mount lookup LOR
When a lookup operation crosses into a new mountpoint, the mountpoint
must first be busied before the root vnode can be locked. When a
filesystem is unmounted, the vnode covered by the mountpoint must
first be locked, and then the busy count for the mountpoint drained.
Ordinarily, these two operations work fine if executed concurrently,
but with a stacked filesystem the root vnode may in fact use the
same lock as the covered vnode. By design, this will always be
the case for unionfs (with either the upper or lower root vnode
depending on mount options), and can also be the case for nullfs
if the target and mount point are the same (which admittedly is
very unlikely in practice).
In this case, we have LOR. The lookup path holds the mountpoint
busy while waiting on what is effectively the covered vnode lock,
while a concurrent unmount holds the covered vnode lock and waits
for the mountpoint's busy count to drain.
Attempt to resolve this LOR by allowing the stacked filesystem
to specify a new flag, VV_CROSSLOCK, on a covered vnode as necessary.
Upon observing this flag, the vfs_lookup() will leave the covered
vnode lock held while crossing into the mountpoint. Employ this flag
for unionfs with the caveat that it can't be used for '-o below' mounts
until other unionfs locking issues are resolved.
John Baldwin [Wed, 26 Oct 2022 21:22:56 +0000 (14:22 -0700)]
vmm: Fix AP startup with old userspace binaries.
Older binaries that do not request IPI exits to userspace do not
start user threads for other vCPUs until a STARTUP IPI triggers a
VM_EXITCODE_SPINUP_AP exit to userland. This means that those vcpus
are not yet active (in terms of vm_active_cpus) when the INIT and
STARTUP IPIs are delivered to the vCPUs.
The changes in commit 0bda8d3e9f7a changed the INIT and STARTUP IPIs
to reuse the existing vlapic_calcdest() function. This function
silently ignores IPIs sent to inactive vCPUs. As a result, when using
an old bhyve binary, the INIT and STARTUP IPIs sent to wakeup APs were
ignored.
To fix, restructure the compat code for the INIT and STARTUP IPIs to
ignore the results of vlapic_calcdest() and manually parse the APIC ID
and resulting vcpuid. As part of this, make the compat code always
conditonal on the ipi_exit capability being disabled.
pkgbase: Put zfs utilities and lib in their own package
It is useful to have zfs utilities and lib in a separate package as
it allow users to create image that can support ZFS (i.e. not with
WITHOUT_ZFS in src.conf set) without bloating the default image with
all zfs tools (for example for jails).
For most users it's not needed to boot and they are also
available in the FreeBSD-rescue package in case an update
break and FreeBSD-geom package isn't updated correctly.
If the system has been up more longer than a minute, we add 30 seconds to
the uptime so that subsequent calculations will round to the nearest minute
rather than truncate. However, since the introduction of libxo, we output
the raw value after performing the adjustment. Rewrite so that we output
the raw value first, then perform the adjustment and recalculate before
outputting the humanized value.
While there, reduce stack usage and avoid needless allocations.
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D37128
Hyper-V: Code refactor to avoid redundancy of MSR values on x86 and arm64
Refactor the code to put split the MSR values for x86 and arm64
Hyper-V. Code not yet built. This is one of several patches for
the arm64 Hyper-V enablement.
Sponsored by: Microsoft
Differential Revision: https://reviews.freebsd.org/D37103
Jose Luis Duran [Thu, 13 Oct 2022 16:11:31 +0000 (13:11 -0300)]
strfmon(3): Fix # explanation
There's only one value that specifies the number of digits after the
decimal point (oh, sorry, the "radix character") the other specifies the
number before...
While here, add a little more info on the effects of using the #n value.
Jose Luis Duran [Fri, 21 Oct 2022 19:34:09 +0000 (16:34 -0300)]
strfmon: Fix formatting of a second fixed-width value
There is a bug when formatting two consecutive values using fixed-widths
and the values need padding. This was because the value of pad_size
was zeroed only every other time.
Format Before After
[%8n] [%8n] [ $123.45] [ $123.45] [ $123.45] [ $123.45]
Jose Luis Duran [Tue, 18 Oct 2022 02:24:03 +0000 (23:24 -0300)]
strfmon: Fix an edge case when sep_by_space is 2
Fix an edge case by printing the required space when, the currency
symbol succeeds the value, a space separates the sign from the value and
the sign position precedes the quantity and the currency symbol.
> When {p,n,int_p,int_n}_sep_by_space is 2:
> If the currency symbol and sign string are adjacent, a space separates
> them; otherwise, a space separates the sign string from the value.
Warner Losh [Tue, 25 Oct 2022 05:05:07 +0000 (23:05 -0600)]
subr_physmem: Fix userspace build
Include stdbool.h in userspace configurations. For the kernel builds we
get it from sys/types.h, but bool isn't defined there for non-kernel
builds and this otherwise kernel-only file is used for the physmem test
suite.
x86/include/elf.h: make inclusion blocks for elf32.h and elf64.h similar
They were copy-pasted when x86/include/elf.h file was merged from its
i386 and amd64 counterparts. Having the text around inclusions
significantly different is somewhat confusing.
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D37085
i386: move hard-coded load address for PIE below default linker base
both for i386 native and compat32 amd64. We know the ld-elf.so.1 size
in advance, it fits there. Trying to push it up after the end of a
binary cannot work reliably and eventually fail for large binaries.
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D37085
arm, arm64: tweak hard-coded load addresses for PIE binaries
They are used when ASLR is not applied.
The need for adjusting is due to rtld direct exec mode puts ld-elf.so.1
at the PIE load address, and this address must not conflict with the
default linker' load address for non-PIE binaries. Otherwise rtld in
direct mode cannot activate image. Example of implicit failure is ldd(1)
refusing to run.
Reported by: kp
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D37085
Warner Losh [Thu, 6 Oct 2022 03:55:26 +0000 (21:55 -0600)]
physmem: Add physmem_excluded to query if a region is excluded
In order to safely reuse excluded memory when it's reserved for special
purpose, we need to test whether or not the memory has been reserved
early in boot. physmem_excluded will return true when the entire range
is excluded, false otherwise.
Warner Losh [Thu, 6 Oct 2022 02:56:43 +0000 (20:56 -0600)]
efi: Add linux memory reserve table defniitions
There is some hardware which can't be completely reset to release the
memory it is using(so far only the GICv3 on arm has fit this
bill). Since that meory needs to be reserved by the OS for that
hardware's later use of it, create defines for code that will parse that
memory table. Otherise the system may allocate the memory for block I/O,
network packets, etc which will lead to memory corruption.
When booting via Linux's kexec protocol, it will add this table to the
EFI systbl's cfgtbl array. While the mechanism to pass 'configuration'
is standardized, these specific tables are not documented except in the
Linux source. Include comments gleened from its study.
Ed Maste [Mon, 24 Oct 2022 18:06:41 +0000 (14:06 -0400)]
build: Use `rm -fv` for BATCH_DELETE_OLD_FILES
It's possible to have files with odd permissions in the tmproot (or
sysroot), causing rm to prompt for each one during e.g. buildworld.
Add -f to forcibly delete these.
Reviewed by: brooks
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D37111
Mark Johnston [Tue, 25 Oct 2022 13:16:23 +0000 (09:16 -0400)]
bhyve: Address warnings in blockif_proc()
- Use unsigned types for all arithmetic. Use a new signed variable for
holding the return value of pread() and pwrite().
- Handle short I/O from pwrite().
Mark Johnston [Fri, 9 Sep 2022 00:40:02 +0000 (20:40 -0400)]
bhyve: Avoid shadowing global variables in bhyverun.c
- Rename the global cores/sockets/threads to cpu_cores/sockets/threads.
This way, num_vcpus_allowed() doesn't shadow them.
- The global maxcpus is unused, remove it for the same reason.
Kyle Evans [Tue, 23 Aug 2022 02:08:03 +0000 (21:08 -0500)]
split: add some tests
This should cover all of the basic functionality, as well as the recent
enhancement to use a dynamic buffer size rather than limiting patterns
and lines to MAXBSIZE.
Reviewed by: bapt
Sponsored by: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D36324
Kyle Evans [Tue, 23 Aug 2022 02:05:58 +0000 (21:05 -0500)]
split: switch to getline() for line/pattern matching
Get rid of split's home-grown logic for growing the buffer; arbitrarily
breaking at LONG_MAX bytes instead of 65536 bytes gives us much more
wiggle room. Additionally, we'll actually fail out entirely if we can't
fit a line, which makes noticing this class of problem much easier.
Mark Johnston [Mon, 24 Oct 2022 21:31:11 +0000 (17:31 -0400)]
libvmmapi: Provide an interface for limiting rights on the device fd
Currently libvmmapi provides a way to get a list of the allowed ioctls
on the vmm device file, so that bhyve can limit rights on the device
file fd. The interface is rather strange: it allocates a copy of the
list but returns a const pointer, so the caller has to cast away the
const in order to free it without aggravating the compiler.
As far as I can see, there's no reason to make a copy of the array, but
changing vm_get_ioctls() to not do that would break compatibility. So
this change just introduces a better interface: move all rights-limiting
logic into libvmmapi.
Any new operations on the fd should be wrapped by libvmmapi, so also
discourage use of vm_get_device_fd(). Currently bhyve uses it only when
limiting rights on the device fd.