Jessica Clarke [Sun, 29 Jan 2023 23:07:25 +0000 (23:07 +0000)]
tools/build/make.py: Drop problematic Linux hack for bmake bootstrap
Since e2eeea75eb8b ("Merge bmake-20201117") missing/sys/cdefs.h has
been present in bmake, and _GNU_SOURCE seems to have been defined by
config.h for much longer than that, possibly for the entire time OS
cross-build support has been in-tree, so these are obsolete. Moreover,
since 79e02149fcb4 ("Fix dtrace tools bootstrap on non-FreeBSD after
OpenZFS import"), HAVE_STRLCAT and HAVE_STRLCPY have been defined by our
cross-build headers in order to placate DTrace tools (which is not the
right way to solve that problem, but motivates fixing this one). Commit 4fde40d9b540 ("Merge/update to bmake-20230126") changed the strlcpy.c in
bmake from including config.h directly to including make.h, which means
it includes string.h and thus sees these bogus definitions, causing it
to not define the strlcpy compat function on Linux even though it needs
to and thus failing to link. Thus, fix this whole mess by removing the
hack we no longer need.
John Grafton [Sun, 29 Jan 2023 23:05:15 +0000 (15:05 -0800)]
rescue: Add fetch(1) to the rescue tool.
After a failed upgrade, having fetch(1) on a system that is
physically unnreachable would be very useful to download files
required to get the OS back up and functional.
On my system this adds 589,824 bytes (3.8%) to the binary size.
* Parse and export newly-added NL_RTA_WEIGHT attribute, providing path
weight for a non-multipath route. This fixes a number of tests in
sys/net/routing which rely on this data.
* Remove handling of NL_RTA_KNH_ID in multipath routes, as it is
not provided.
* Improve kernel/user nexthop index export. As a result,
for multipath routes:
* nhg-kidx attribute represents kernel nhg index (always provided)
* nhg-uidx attribute represents user-provided nhg index (if set)
for non-multipath routes:
* nhop-kidx attribute represents kernel nhop index (always provided)
* nhop-udx attribute represents user-provided nexthop index (if set)
netlink: allow path weight manipulations for single-path routes.
Add support for the scenario when user adds/deletes paths for a single
prefix one-by-one, all with different weights.
This change adds a new FreeBSD-specific RTA attribute, NL_RTA_WEIGHT.
When dumping non-multipath routes, this attribute is added if the
route weight is not RT_DEFAULT_WEIGHT.
When adding a new route, this attribute is parsed as a relative path
weight.
Andrew Turner [Sat, 4 Jun 2022 11:13:51 +0000 (12:13 +0100)]
Make SMCCC usable by device drivers
To allow device drivers to call into SMCCC we need to initialise it
earlier. As it depends on PSCI, and that is detected via ACPI or FDT
move the call to smccc_init to the PSCI driver.
Add a function for drivers to read the smccc version, or 0 if smccc
is not present.
Andrew Turner [Sat, 28 Jan 2023 17:36:24 +0000 (17:36 +0000)]
Disable the arm physical timer when an irq exists
Some firmware leaves the timers enabled. Ensure they are disabled if
there are any physical timer interrupt resources to ensure we don't
receive any unexpected interrupts from them.
Dmitry Chagin [Sun, 29 Jan 2023 13:17:33 +0000 (16:17 +0300)]
cpuset: Fix sched_[g|s]etaffinity() for better compatibility with Linux.
Under Linux to sched_[g|s]etaffinity() functions the value returned from a call
to gettid(2) (thread id) can be passed in the argument pid. Specifying pid as 0
will set the attribute for the calling thread, and passing the value returned
from a call to getpid(2) (process id) will set the attribute for the main thread
of the thread group.
Native cpuset(2) family of system calls has "which" argument to determine how
the value of id argument is interpreted, i.e., CPU_WHICH_TID is used to pass
a thread id and CPU_WHICH_PID - to pass a process id.
For now native sched_[g|s]etaffinity() implementation is wrong as uses "which"
CPU_WHICH_PID to pass both (process and thread id) to the kernel. To fix this
adding a new "which" CPU_WHICH_TIDPID intended to handle both id's.
Dmitry Chagin [Sun, 29 Jan 2023 12:36:25 +0000 (15:36 +0300)]
libthr: Fix pthread_[g|s]etaffinity_np manual.
Since f35093f8 semantics of a thread affinity functions is changed to be a
compatible with Linux:
In case of getaffinity(), the minimum cpuset_t size that the kernel permits is
the maximum CPU id, present in the system, / NBBY bytes, the maximum size is not
limited.
In case of setaffinity(), the kernel does not limit the size of the user-provided
cpuset_t, internally using only the meaningful part of the set, where the upper
bound is the maximum CPU id, present in the system, no larger than the size of
the kernel cpuset_t.
Dmitry Chagin [Sun, 29 Jan 2023 12:35:18 +0000 (15:35 +0300)]
libthr: Fix pthread_attr_[g|s]etaffinity_np to match it's manual and the kernel.
Since f35093f8 semantics of a thread affinity functions is changed to be a
compatible with Linux:
In case of getaffinity(), the minimum cpuset_t size that the kernel permits is
the maximum CPU id, present in the system, / NBBY bytes, the maximum size is not
limited.
In case of setaffinity(), the kernel does not limit the size of the user-provided
cpuset_t, internally using only the meaningful part of the set, where the upper
bound is the maximum CPU id, present in the system, no larger than the size of
the kernel cpuset_t.
To match pthread_attr_[g|s]etaffinity_np checks of the user-provided cpusets to
the kernel behavior export the minimum cpuset_t size allowed by running kernel
via new sysctl kern.sched.cpusetsizemin and use it in checks.
Dmitry Chagin [Sun, 29 Jan 2023 12:34:10 +0000 (15:34 +0300)]
sched.h: Fix _S macros for better compatibility with glibc.
In e2650af157 was added "_S" macros for compatibility with glibc, but it's still
incompatible as under glibc the macros whose names end with "_S" operate on the
dynamically allocated CPU set(s) whose size is in bytes, not in bits.
While here remove limiting ifdef to non-kernel case.
Fedor Uporov [Mon, 5 Dec 2022 17:04:42 +0000 (20:04 +0300)]
Add extended attributes
The extattrs follows semantic of ufs, mean it cannot
be set to char/block devices and fifos. The attributes
are allocated using regular malloc with M_WAITOK
allocation with the own malloc tag M_TMPFSEA. The memory
consumed by extended attributes is limited to avoid OOM
triggereing by tmpfs_mount variable tm_ea_memory_max,
which is set initialy to 16 MB. The extended attributes
entries are stored as linked list in the tmpfs node.
The mount point lock is required only under setextattr
and deleteextattr to update extended attributes
memory-inuse counter, all other operations are doing
under vnode lock.
Fedor Uporov [Thu, 26 Jan 2023 10:23:22 +0000 (13:23 +0300)]
Fix pjfstest issue tests/rename/23.t
This test creates two files like file0 and file1,
then creates link to file1 and checks ctime on it.
Then renames file0 to file1. Then checks ctime on
link again. It is expected, that second ctime will
be higher then first ctime, because rename happen.
Add ctime updating for directory entry,
which will be deleted on rename.
Fedor Uporov [Thu, 26 Jan 2023 10:17:48 +0000 (13:17 +0300)]
Fix pjfstest issue tests/rename/19.t
The rename call with args like:
"./dir0/dir1/.." "./dir2" will cause MPASS failure.
The tmpfs_dir_lookup() does not accept names like
'.' and '..' for lookup. Move the '.' and '..' entry
check before tmpfs_dir_lookup() call.
Stefan Eßer [Sat, 28 Jan 2023 22:48:34 +0000 (23:48 +0100)]
usr.bin/gh-bc: update Makefiles for version 6.2.2
The sources of this program are in contrib/bc, but built using the
Makefiles touched in this commit, which had to be adapted to comply
with changed made to the build system of this software, which is
not used when building in the base system.
Stefan Eßer [Sat, 28 Jan 2023 21:26:22 +0000 (22:26 +0100)]
contrib/bc: merge from vendor release 6.2.2
This update fixes a few issues in history editing and the processing
of the "quit" function. The "quit" function will no longer cause bc
to exit when encountered in a script file (before any command from
the script has been executed).
New functions is_number(), is_string return 1 if the passed argument
is a number resp. a string. The asciify() function has been extended
to support the conversion of an array of numbers into a string.
Stefan Eßer [Sat, 28 Jan 2023 20:08:43 +0000 (21:08 +0100)]
vendor/bc: import version 6.2.2
This is a production release that fixes a bug.
The bug was that if an array element was used as a parameter, and then
a later parameter had the same name as the array whose element was
used, bc would grab the element from the new array parameter, not the
actual element from before the function call.
Stefan Eßer [Sat, 28 Jan 2023 20:02:27 +0000 (21:02 +0100)]
vendor/bc: import version 6.2.0
This is a production release with a new feature and a few bug fixes.
The bug fixes include:
- A crash when bc and dc are built using editline, but history is not
activated.
- A missing local in the uint*() family of functions in the extended
math library.
- A failure to clear the tail call list in dc on error.
- A crash when attempting to swap characters in command-line history
when no characters exist.
- SIGWINCH was activated even when history was not.
The new feature is that stack traces are now given for runtime errors.
In debug mode, the C source file and line of errors are given as well.
Stefan Eßer [Sat, 28 Jan 2023 19:59:18 +0000 (20:59 +0100)]
vendor/bc: import version 6.1.0
This is a production release that fixes a discrepancy from the bc
standard, a couple of memory bugs, and adds new features.
The discrepancy from the bc standard was with regards to the behavior
of the quit command. This bc used to quit whenever it encountered quit
during parsing, even if it was parsing a full file. Now, bc only quits
when encountering quit after it has executed all executable statements
up to that point.
This behavior is slightly different from GNU bc, but users will only
notice the difference if they put quit on the same line as other
statements.
The first memory bug could be reproduced by assigning a string to a
non-local variable in a function, then redefining the function with
use of the same non-local variable, which would still refer to a
string in the previous version of the function.
The second memory bug was caused by passing an array argument to the
asciify() built-in function. In certain cases, that was wrongly
allowed, and the interpreter just assumed everything was correct and
accessed memory. Now that arrays are allowed as arguments (see below),
this is not an issue.
The first feature was the addition of the is_number() built-in
function (u in dc) that returns 1 if the runtime argument is a number
and 0 otherwise.
The second feature was the addition of the is_string() built-in
function (t in dc) that returns 1 if the runtime argument is a string
and 0 otherwise.
These features were added because I realized that type-checking is
necessary now that strings can be assigned to variables in bc and
because they've always been assignable to variables in dc.
The last added feature is the ability of the asciify() built-in
function in bc to convert a full array of numbers into a string. This
means that character-by-character printing will not be necessary, and
more strings than just single-character ones will be able to be
created.
Dmitry Chagin [Sat, 28 Jan 2023 10:19:41 +0000 (13:19 +0300)]
linux(4): Rework statfs conversion routine.
Rework the routines to convert a native statfs structure (with fixed-size 64-bit
counters) to a Linux statfs structure (with long-sized counters) for 32-bit apps.
Instead of following Linux and return an EOVERFLOW error from statfs() family of
syscalls when actual fs stat value(s) are large enough to not fit into 32 bits,
apply scale logics used by FreeBSD to convert a 5.x statfs structure to a 4.x
statfs structure.
Relevant/interesting changes (see ChangeLog for more):
o variables like .newline and .MAKE.{GID,PID,PPID,UID}
should be read-only.
o .[NO]READONLY: for control of read-only variables
o .SYSPATH: for controlling the path searched for makefiles
o allow for white-space between command specifiers @+-
o add more details to warning 'Extra targets ignored'
o make.1: sync list of built-in variables with reality
sort list of built-in variables
o cond.c: add more details to error message for numeric comparison
o job.c: fix handling of null bytes in output
o Allow .break to terminate a .for loop early
o var.c: fix out-of-bounds errors when parsing
o fix exit status for '-q' (since 1994)
Mitchell Horne [Fri, 27 Jan 2023 22:00:07 +0000 (18:00 -0400)]
rtalloc.9: remove obsolete man page
This KPI was removed in d22337254564. Note that there are a handful of
references remaining in the src tree to these rtalloc functions that
could be cleaned up by someone with more domain knowledge.
Reviewed by: pauamma (manpages), glebius, melifaro
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D38188
linuxkpi: Include `linux/sched/mm.h` from `linux/sched.h`
At least one file in the DRM drivers benefits from some namespace
pollution to use `fs_reclaim_acquire()`/`fs_reclaim_release()`. They are
defined in `linux/sched/mm.h` and this header must be included
indirectly into the DRM drivers' source file.
I couldn't find how it was included. Therefore this commit includes
`linux/sched/mm.h` from `linux/sched.h`. This is not the case in Linux
but fixes the issue with the DRM drivers.
Reviewed by: manu
Approved by: manu
Differential Revision: https://reviews.freebsd.org/D37912
amd64: do not enable PKRU if user disabled saving PKRU register in xsave mask
This is done by reverting CR4_PKE bit, because we perform %CR4
initialization in initializecpu(), and the function is called before
xsave_mask is read. To not redo the whole early initialization
sequence for the corner case, this should be good enough.
Reported by: jhb
Reviewed by: jhb, markj
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Differential revision: https://reviews.freebsd.org/D38219
netstat: make netstat -rn use netlink instead of rtsock
This change switches route listing in netstat to netlink, with fallback to rtsock.
The outputs are mostly identical, with an exception of not showing kernel
nexthop indexes for multipath routes.
Gleb Smirnoff [Fri, 27 Jan 2023 15:09:22 +0000 (07:09 -0800)]
sysctl: use correct types and names in sysctl_*sec_to_sbintime
The functions are intended to report kernel variables that are
stored as sbintime_t (pointed to by arg1) as human readable
nanoseconds or milliseconds (reported via sysctl_handle_64).
The variable types and names were reversed. I guess there is
no functional change here, as all types flipped around were
signed 64. Note that these function aren't used yet anywhere
in the kernel.
Ed Maste [Wed, 16 Nov 2022 14:20:39 +0000 (09:20 -0500)]
Retire WITHOUT_CXX option
Several important base system components are written in C++, and the
WITHOUT_CXX option produced a system that was not fully functional.
Just accept this, and remove the option to build without C++ support.
Ed Maste [Wed, 25 Jan 2023 22:01:00 +0000 (17:01 -0500)]
lua: reduce diffs between luaconf.h copies
Upstream luaconf.h is contrib/lua/src/luaconf.h.dist, while userland lua
and loader lua have copies in lib/liblua/luaconf.h and
stand/liblua/luaconf.h.
Adjust whitespace, VCS tags, etc. to match upstream's version, for ease
of comparison.
Reviewed By: imp
Sponsored By: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D38206
During tcp session start, various mechanisms need to
track a few initial RTTs before becoming active.
Prevent overflows of the corresponding tracking counter
and reduce the size of tcpcb simultaneously.
Mark Johnston [Thu, 26 Jan 2023 15:46:19 +0000 (10:46 -0500)]
netlink: Zero-initialize writer structures allocated on the stack
The prevailing pattern seems to be to simply initialize all fields to
zero. Without this, it's possible to trigger a branch on uninitialized
memory, specifically, when testing nw->ignore_limit in
nlmsg_refill_buffer().
Initialize the writer structure in a couple of functions where this is
necessary.
Reported by: KMSAN
Reviewed by: melifaro
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D38213
This is the same as `debugfs_create_file()` but takes the initial size
of the file. In FreeBSD, the given size is ignored and
`debugfs_create_file()` is called.
Reviewed by: emaste, manu
Approved by: manu
Differential Revision: https://reviews.freebsd.org/D37914
By (ab)using `atomic_long_add_return()`, `atomic_long_sub()` was making
the atomic long overflow. Indeed the underlying FreeBSD atomic is based
on an unsigned long.
Reviewed by: manu
Approved by: manu
Differential Revision: https://reviews.freebsd.org/D38090
However, these commits were later reverted (also in Linux 5.13) so the
i915 driver doesn't use these functions. But perhaps it will help in the
future.
To sum up, the code comes from the i915 DRM driver but it doesn't use it
(i.e. it continues to use its internal implementation).
Reviewed by: manu
Approved by: manu
Differential Revision: https://reviews.freebsd.org/D38088
projects / FreeBSD / FreeBSD.git / log