There are now some system headers (e.g. <stdatomic.h>) that have outright
different versions for C and C++ which are not compatible with each other.
Change-Id: Ibf797e0817f0fe4d5241424d7d06023b19888c02
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6991
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Introduce new attribute classes:
* addrptr represents an offset into the .debug_addr_section
* loclist represents an offset or an index to a location list
* rangelist represents an offset or an index to a range list
* stroffsetsptr represents an index to the indirect string table
The semantics of some existing attribute classes changed:
* loclistptr represents an offset into the
.debug_loc / .debug_loclist section
It will be used only when setting loclists_base attribute.
In all other places we should use loclist attribute class.
* rangelistptr represents an offset into the
.debug_ranges / .debug_rnglist section
It will be used only when setting rnglists_base attribute.
In all other places we should use rangelist attribute class.
Define lookup table for new DW_AT_* attribute types
* change the existing references to AC_LOCLISTPTR to AC_LOCLIST
as in those places now either an index or an offset can be accepted
(previously there were no indexes, only offsets everywhere)
* only DW_AT_loclists_base uses AC_LOCLISTPTR
* similarly for AC_RANGELISTPTR and AC_RANGELIST as ranges
can be specified using either an index or an offset
* only DW_AT_rnglists_base uses AC_RANGELISTPTR
Define lookup table for new DW_FORM_* attribute forms
* similar reshuffle for location lists and range lists
Change-Id: I78d307889cd6082a19870afdc571dc0c83d19644
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6982
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This undoes fc7864091e2463c0e446f8b3954df4d29e9465af.
It was only ever used by the FreeBSD compatibility layer,
only supported one structure, and created problems with
concurrent builds.
Use sizeof()+roundup() to build a char[] of the correct size,
and add a static_assert to ensure it stays so.
The <kernel_c++_structs.h> header needs to have its dependency
manually declared, and not all consumers of this file properly
declared it as such. This then fixes a concurrent build problem.
Fixes #17965 as this was the only consumer of this header.
Someone asked about it in the forum.
Change-Id: Ib80b9a3bd76f6e882070cc60fcc239f5837d8231
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6983
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
If a pinned thread is being run while disabling a CPU core, the thread
is properly requeued to that core, but with fCore set to NULL, causing
the scheduler to panic the next time that thread is rescheduled.
Change-Id: I9cf18c5a470517124b2573347cf3eb6fd26ab7b6
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6917
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
* introduce new field unitType. Currently we support only
DW_UT_compile unit type.
* abbrevOffset and addressSize are shuffled around
Change-Id: I67db242240012b52e0ec59f87ffab7cb937eb212
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6981
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
For attributes with form DW_FORM_implicit_const, the attribute
specification contains a third part which is a signed LEB128
number. This value shall be used as the value of the attribute
when parsing the debug_info section.
Change-Id: I3e9aa3ba70b3aa05e6d2f570a1464966c38e4b1b
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6980
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Rene Gollent <rene@gollent.com>
The 'whence' function defined in /etc/profile serves a limited purpose
with Bash as the default shell. Bash already provides
which/type/command as built-ins, which the function simply wraps.
Additionally, zsh has a built-in 'whence' command, which is masked by
the definition in /etc/profile. The function does not work under zsh,
breaking core zsh functionality as a result.
Fixes #18612
Change-Id: Ia34d95ffd58b2cb06d71145a217a9017657fa4a3
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6979
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
* add new field for .debug_line_str section
* implement reading string and int values tagged with DW_FORM_*
* skip new fields addressSize, segmentSelectorSize in line info header
* parse directories and file names according to new syntax
* enable line-info parsing up to version 5
Change-Id: I1dd7e0d834c73f6d72b7115c7ded9f868be84a3d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6978
Reviewed-by: Rene Gollent <rene@gollent.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
The structure of a logical APIC destination is made up of a 16-bit
bitfield of cluster-local IDs, as well as a 16-bit cluster ID. When
sending an ICI to multiple cores in different clusters, these cannot simply be ORed together.
Fixes #17233
Change-Id: Ifa84da51eccd85c1eff529749ffa00bc2159899e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6919
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
R_68K_JMP_SLOT should be handled similarly to R_68K_GLOB_DAT i.e.
the relocated value is symbol_value + addend. see e.g. in uclibc:
https://git.uclibc.org/uClibc/tree/ldso/ldso/m68k/elfinterp.c#n222
(also the G variable is never assigned any nonzero value?)
before this change we get jumps to 0x00000000, bus errors with
pc=0x00000030 etc.
after this change we are able to call functions!
Change-Id: Ibe135adb0dc6248a2baf5d6eca1543fdddf3a9e8
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6957
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Fixes some PTE concurrent access bugs.
Change-Id: I09ec56861fae389a8a3e228b17a3921b85202c8b
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6949
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Attribute is ignored for now.
It is supposed to check ABI compatiblity and reject loading incompatible images.
Haiku currently do not use multiple ABIs for RISC-V so it is safe to ignore attributes.
PT_RISCV_ATTRIBUTES program header is produced by default in GCC 13 and Clang 17.
Change-Id: I4659e9bacbf34a2a0bc16b34c2aaa37232d700fa
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6948
Reviewed-by: David Karoly <karolyd577@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
This error only occurs when compiling a debug build. It was introduced
with commit fb885767b6f
Change-Id: I80b470df3872b5bdd8690f65b3d14388158be8ed
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6951
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This switches the application over from using basic
authentication to using token-based authentication in
preparation for later using Open-ID based
authentication flows. The application version is also
bumped in order that the server can detect this version
at some later date in the future when it no longer
supports basic authentication itself.
Change-Id: I7addde1d57503c58d6bcd54908f22f66830c0c59
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6944
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
The comment above this block specified the correct behavior: the
button's status is already updated elsewhere, so just check IsEnabled().
But the code did not actually do that.
It appears this has been incorrect all the way back to the initial
OpenTracker import in 2001.
Fixes #13721.
In hrev51155, it was changed to be always enabled no matter what,
so that the current directory (without anything selected) could be opened.
But this means it would also be enabled for files, which is misleading.
Instead, remove the function, and adjust the logic detecting whether
to enable the button or not to take B_DIRECTORY_NODE into account.
CVS revision(s) are in first-line comments.
Requires a new firmware package (from HaikuPorts). There are some
changes in here which may fix some of the odder issues, like failures
to connect to access points. There is also support for one new PCI ID.
Tested by PulkoMandy who reports it's still working.
* Replace count_low/count_high with bigtime_t fields plus an int32.
sizeof(spinlock) is now 32 bytes with the debug option enabled.
* Adjust and clean up all spinlock code to use the new fields.
* Fold DEBUG_SPINLOCK_LATENCIES into the new code. Remove the bootloader
option and other flags for it (these were not compiled in by default.)
The new code should be much easier to understand and also more powerful.
However, the information transmitted to userland isn't as useful now;
the KDL command output will have the interesting information.
(Things could be reworked to transmit more interesting information to
userland again if desired, but as this code clearly hadn't been compiled
for many years, as it referred to global spinlocks that have been gone
for a very long time.)
Change-Id: I2cb34078bfdc7604f288a297b6cd1aa7ff9cc512
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6943
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Only the x86 and PPC implementations look like they would have worked,
while the builtin is available and will work across all architectures.
We already use it unconditionally in some parts of libroot.
Change-Id: I2dffb3b2c7cdd605092382b9d649151adb921bb4
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6942
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This is the loop that runs on all CPUs besides the one
actually running the kernel debugger. It's functionally
a spin-loop around a few different variables, and so it
should be safe to use cpu_pause() here, just as real
spinlocks do.
(cpu_pause() just runs a "pause" or equivalent instruction,
it doesn't use the CPU-idle modules, which indeed may be
unsafe to use in KDL.)
A glance at FreeBSD seems to indicate they also do this
when their kernel debugger is active.
Seriously improves power consumption while KDL is active:
even running in a VM with only a few cores, there would be
obvious fan spin-up when entering KDL. After this change,
there's barely any at all (while overall CPU usage % remains
basically identical.)
I looked at this code as a reference when writing similar code in another
project, and realized then that this was broken to the point where I
don't think anyone could have actually used it:
1. Writes larger than the buffer should only return the amount written
by this Write() call, not the amount flushed beforehand.
2. fDirty was only set if there were bytes remaining after the first
write to the buffer. So, if you wrote small amounts of data, they
would be silently discarded most of the time!
3. When filling the buffer, we might as well flush simultaneously.
This allows the logic to be consolidated into a loop and remove the
duplicate memcpy. In case of failure, just return the bytes-written
as the write could be retried later.
4. Flush() should always return an error or 0, not bytes written.
While at it, add some basic tests for this class.
Change-Id: I2de01d0b31e3fe22863cef21dd7b0b62ed47121b
* Remove unused/unneeded parameters to Dequeue.
* Make use of StackOrHeapArray.
* Reorder syscall-entry checks for efficiency.
* Inline the unlock-block method and unset variables in the process.
* Reorder code for clarity and to reduce indentation.
When breaking out of _RemovePackageContentRootNode() by encountering
a specified "endPackageNode", we need to unlock all directories.
This case is only possible to encounter when _AddPackageContentRootNode()
fails partway and invokes _Remove() to remove the just-added nodes.
Fixes a hang/deadlock encountered by Diver while testing some (disabled)
HaikuPorts recipes.
