The manually written code was all wrong (missing branch delay slots,
wrong type of return instruction used, probably more bugs). Use the same
approach as x86 to have inline functions instead, which is much better
for performance and simpler to write.
Change-Id: Iac0fc814c15311658f983da58ac7f9d3edd75b81
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3595
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
The definition in SupportDefs.h using gcc builtins is sufficient. No
need for a custom one. The same approach is used on x86 with gcc8
already, but other platforms had not been adjusted to use it.
Change-Id: I3973ff723a31f90cc8d19ac098eb1e85d471d610
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3594
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
The sparc ABI reserves the g7 register for this.
Change-Id: I93b81ecef72cde859972ef7b7f6b9991d35f9f29
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3583
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
It was bumped for bios and efi from previously very low values, but
other architectures did not follow.
Change-Id: I6ce92e2cdb0261d4d0637753e77d555d407073fc
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3575
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
* Makes our UEFI bootloader somewhat FDT/DTB aware on all
architectures.
* Will report when an FDT is found, and provide it to kernels
that want it.
Change-Id: I90324fc0579a9c835e60568fa9b654c2df0aba27
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3543
Reviewed-by: Fredrik Holmqvist <fredrik.holmqvist@gmail.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
the preallocate syscall will call the preallocate filesystem hook, if available.
fix#6285
Change-Id: Ifff4595548610c8e009d4e5ffb64c37e0884e62d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3382
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Currently used by fixup_next_boot_floppy.
Change-Id: I47c10657b5280f00e470a3171ad11744859ce76c
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3310
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
when the cpufreq module is loaded, we let the scheduler update its policy.
Improve assert report
CoreEntry::GetLoad() could return more than kMaxLoad.
Change-Id: I127f9b3e8062b5996872aae30b4021b9904fa179
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3216
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
implement on x86 with APERFMPERF.
Change-Id: Ia484854c76dee76c5447983de15800a25d791d39
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3213
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
to call a function on the target cpu. Early mechanism not available.
Change-Id: I9d049e618c319c59729d1ab53fb313b748f82315
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3212
Reviewed-by: Axel Dörfler <axeld@pinc-software.de>
When a thread is created, it is expected that some other thread (usually the
creating thread) will want to make sure it completes. This is done using the
pthread_join() or wait_for_thread() calls.
It is possible that threads end before another thread waits for its completion.
That's why there is a dead thread list for each team, which holds thread ids
and their exit status so that a call to pthread_join() or wait_for_thread() in
the future can complete succesfully.
The dead thread list was limited to 32 threads per team. If there would be
more, the oldest thread would be kicked off. This could cause issues in
situations where a team would create more than 32 threads, and would start
waiting for their result after they have finished. Some of the calls would fail
because the threads would no longer be in the dead list.
This specifically caused problems for cargo (the Rust package manager), which
could depending on the number of dependencies, could create more than 32
threads. See: https://github.com/nielx/rust/issues/3
This change removes the limit of dead threads within a team. Note that there is
a risk that a badly written program that does not detach or joins its threads
can make this an endless list, but the impact is relatively small (dead threads
only occupy a bit of kernel memory).
Change-Id: I0135dd54e10ee48a529f23228d21237d4f1a74e2
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3178
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
When copying an area with vm_copy_area only the new protection would be
applied and any possibly existing page protections on the source area
were ignored.
For areas with stricter area protection than page protection, this lead
to faults when accessing the copy. In the opposite case it lead to too
relaxed protection. The currently only user of vm_copy_area is
fork_team which goes through all areas of the parent and copies them to
the new team. Hence page protections were ignored on all forked teams.
Remove the protection argument and instead always carry over the source
area protection and duplicate the page protections when present.
Also make sure to take the page protections into account for deciding
whether or not the copy is writable and therefore needs to have copy on
write semantics.
Change-Id: I52f295f2aaa66e31b4900b754343b3be9a19ba30
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3166
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Change-Id: Id050fad59ede444f2eab7eca681c6ec44612aaf9
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3160
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: François Revol <revol@free.fr>
This enables generation of exceptions that are due to uncorrected
hardware errors. The exception handlers were already in place and will
now actually trigger kernel panics.
Note that this is the simplest form of MCE "handling" and does not add
anything of the broader machine check architecture (MCA) that also allow
reporting of corrected errors. As MCEs are generally hard to decode due
to their hardware specifity, this merely makes such problems more
obvious.
Might help to discern hardware issues in cases that would otherwise just
triple fault and cause a reboot.
Change-Id: I9e3a2640458f7c562066478d0ca90e3a46c3a325
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3155
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Reviewed-by: Axel Dörfler <axeld@pinc-software.de>
Pages in the given range are unmapped and freed without getting written
back anywhere. It can be used whenever a caller does not care about the
data in the given range anymore and wants to reduce page pressure.
Change-Id: I8bcce68fab278efef710d3714677e1d463504a56
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2843
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
* Migrate some platform agnostic architecture code into
boot/arch from efi/arch. This helps to avoid conflicts
between kernel and boot sources as well.
* Conflicts between arch_cpu in efi and kernel code means
bootcode really should *never* directly use kernel arch
headers. (other platforms don't, which is why they don't
have this same issue)
* We carefully thread any needed kernel headers (namely
assembly helper macros) into the bootloader headers without
mixing in the whole conflicting kernel/arch headers.
* ARM now properly get its cpu init code called, and we
progress further into the EFI bootloader.
Change-Id: If67ec9758b5ce68563ebd9eb45d5196401911c67
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2975
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
None of these were used; they were all imported with the original
root Haiku commit, and they are totally unrelated to PulkoMandy's
new SPARC work. Plus, they were also under a BSD Advertising Clause
license.
xsave or xsavec are supported.
breaks vregs compatibility.
change the thread structure object cache alignment to 64
the xsave fpu_state size isn't defined, it is for instance 832 here, thus I picked 1024.
Change-Id: I4a0cab0bc42c1d37f24dcafb8259f8ff24a330d2
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2849
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
It iterates over all areas intersecting a given address range and
removes the need for manually skipping uninteresting initial areas. It
uses VMAddressSpace::FindClosestArea() to efficiently find the starting
area.
This speeds up the two iterations in unmap_address_range and one in
wait_if_address_range_is_wired and resolves a TODO in the latter hinting
at such a solution.
Change-Id: Iba1d39942db4e4b27e17706be194496f9d4279ed
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2841
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This introduces VMAddressSpace::FindClosestArea() that can be used to
find the closest area to a given address in either direction. This is
now trivial and efficient since both kernel and user address spaces use
a binary search tree.
Using FindClosestArea() getting multiple area infos is sped up
dramatically as it removes the need for a linear search from the first
area to the one given in the cookie on each successive invocation.
Change-Id: I227da87d915f6f3d3ef88bfeb6be5d4c97c3baaa
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2840
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
They return the left and right most nodes of the entire tree, i.e.
starting from the root node.
Change-Id: I651a9db6d12308aef4c2ed71484958428e58c9bc
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2838
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
The code in the Resize and Rebase methods was identical except for the
iterator.
Change-Id: I9f6b3c2c09af0c26778215bd627fed030c4d46f1
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2835
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This allows switching from another recursive_lock, mutex or read-locked
rw_lock analogous to the switching possibilities already in mutex.
With this, recursive_locks can be used in more complex situations where
previously only mutexes would work.
Also add debugger command to dump a recursive_lock.
Change-Id: Ibeeae1b42c543d925dec61a3b257e1f3df7f8934
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2834
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
The memory map may be unordered and include overlapping ranges. To make
sure that nothing gets included as usable that should actually be
excluded, first scan for all usable ranges and add them, then remove
anything unusable from these ranges again.
To calculate the amount of unusable memory, count the total after the
first pass and then subtract the total after the second. This way, only
unusable ranges that actually overlap physical memory (and therefore
reduce the amount of usable memory) get excluded.
Note that the explicit ignore of the ACPI reclaim memory is subsumed by
the above. We still don't want to add this region to the usable memory
map, as that would allow the kernel to allocate pages into that region,
possibly corrupting ACPI tables before they were used. We also don't
want to add it as an allocated range, as it is not guaranteed that ACPI
is done with the tables before the unused bootloader ranges are freed in
the kernel.
Also add the missing unusable memory amount from ignoring the first MiB
of memory in the EFI loader.
May fix#16056 although it is not certain that graphics memory ranges
are actually included in the memory map.
Change-Id: Ie7991d2c4dcd988edac2995b3a7efc509fa0f4a3
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2814
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
I forgot to change MUTEX_INITIALIZER following removal of the
unused field.
Change-Id: I011c023ae00bb4576c8bcecf83546892fef3a77e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2719
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
As far as I can tell, there is no reason to ignore unlocks, ever;
if no threads are waiting, then mutex_unlock() will act appropriately.
So all we need to do is increment the lock's count here,
as we are relinquishing our request for locking.
On the other hand, if we did not find our structure in the lock,
that means we own the lock; so to return with an error from here
without changing the count would result in a deadlock, as the lock
would then be ours, despite our error code implying otherwise.
Additionally, take care of part of the case where we have woken up
by mutex_destroy(), by setting thread to NULL and checking for it
in that case. There is still a race here, however.
May fix#16044, as it appears there is a case where ACPICA
calls this with a timeout of 0 (we should make this be
a mutex_trylock, anyway.)
Change-Id: I98215df218514c70ac1922bc3a6f10e01087e44b
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2716
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Rename MovePageRange to Adopt and group it with Resize/Rebase as it
covers the third, middle cut case.
Implement VMAnonymousCache::Adopt() to actually adopt swap pages. This
has to recreate swap blocks instead of taking them over from the source
cache as the cut offset or base offset between the caches may not be
swap block aligned. This means that adoption may fail due to memory
shortage in allocating the swap blocks.
For the middle cut case it is therefore now possible to have the adopt
fail in which case the previous cache restore logic is applied. Since
the readoption of the pages from the second cache can fail for the same
reason, there is a slight chance that we can't restore and lose pages.
For now, just panic in such a case and add a TODO to free memory and
retry.
Change-Id: I9a661f00c8f03bbbea2fe6dee90371c68d7951e6
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2588
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
* Adds VMCache::MovePageRange() and VMCache::Rebase() to facilitate
this.
Applied on top of hrev45098 and rebased with the hrev45564 page_num_t to
off_t change included.
Change-Id: Ie61bf43696783e3376fb4144ddced3781aa092ba
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2581
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Change-Id: Id43bfcbfc24b1adb8f6e9fff587c6df9b62910f2
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2413
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Previous version of the patch was broken by the EFI refactoring.
Change-Id: I6dd125100b22b2461c531bfd8f81b3dd28e2b751
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2409
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
It is no longer an error to destroy a ConditionVariableEntry
that is still attached to a ConditionVariable; it will
now be implicitly detached in that case.
This makes ConditionVariableEntrys much eaiser to use
from an API standpoint.
Change-Id: I03c676d3a198aa885de733d3e1729b15f80de031
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2301
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Change-Id: I0d6d2f8db2bc86c08d5ba2648f1cf46d85b54a5e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2267
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
* arm efi additions
* cleanup some cpu headers which were oddly
split between efi and bios_ia32
* Move calculate_cpu_conversion_factor over to
arch_timer since it is timerish, and x86 only
* Drop some duplicated code from efi start. Move
hpet init code into efi timer/hpet code
Change-Id: Ia4264a5690ba8c09417b06788febc4f572f111ce
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2259
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
* This is the bulk of the work. Anything else should be
minor cleanups and tweaking.
* riscv64 isn't a viable EFI platform yet.. just acting
as a stand-in to test a non-x86 EFI haiku_loader
Change-Id: Ib03de81e2b562e693987b86d7b4318209fb1c792
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2256
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
On modern x86, one can use __rdtscp to get the current cpu in userland.
Change-Id: I1767e379606230a75e4622637c7a5aed9cdf9ab0
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2248
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
The hacks to still use actual system headers for zlib didn't quite work.
- Define Z_SOLO, which makes zlib build without any system include
- Remove use of std::max and #include <algorithm> from AVLTree
- Do not include DebugSupport.h because it uses system headers
- Do not include uuid.h and define just what we need
Now it's possible to compile the btrfs_shell on Linux.
Change-Id: I74a14b5f6804db45ab5a9f582ab493d696376fd3
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2098
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
This issue was initially detected by PVS Studio (issue number V547) and fixed
as part of Google Code-in 2019.
The initial problem was the calculate_cpu_conversion_factor function
which had been copied in the BIOS and EFI versions of the boot code.
Further investigation led to more duplicated or very similar functions
being identified.
Introduce an arch_cpu.h for the x86 boot platform to group these things
in a single place, and adjust the BIOS and EFI code to call into that.
Note that the BIOS and EFI code is still a little platform specific,
ideally there should be a boot_arch_cpu_init() function for each
architecture as already done for openfirmware and u-boot.
Also remove some irrelevant comments from copypasted files for other
architectures, as that was filling my git grep with useless noise.
Change-Id: I16d815f0bf015cec0b4e03cc14f3cc447c7164c5
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1985
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Get enough of the mmu working to be able to allocate memory.
Unlike on PowerPC, we get both address and size as 64bit values. So
adjust of_region to allow this.
Also unlike the PPC port, we do not drive the hardware directly, instead we
rely on the openboot primitives to manage the translation table. This
allows staying independant of the hardware, which is a good idea at
least for the bootloader (we can do actual hardware things in the
kernel)
Change-Id: Ifa57619d3a09b8f707e1f8640d8b4f71bb717e2a
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1482
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Gets call-method working for sparc, and fix more places where we
accidentally truncate 64bit values or sign-extend 32 bit ones.
Change-Id: Ic79c55ffa8d2b475858def1639004412f17dd0c1
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1986
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
For #15515
As mentionned in the ticket, we may also want to hide the symbols
altogether from libroot for newer API/ABI versions, unless we still want
to provide C89/C99/C++98/C++11 compatibility, in which case we still
need them around.
Change-Id: I0ee267fb6c4c2f4bae9b1ba6f68e2bcefc399a7f
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2061
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
* Drop gnu-efi
Change-Id: Ib601fc8ced49b18281b6b98cf861a5aef1b9c065
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2026
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
* Move x86 hpet behind timer interface.
* Add a few if x86,x86_64 macros to start.cpp.
Change-Id: I583ec1b064785182e6d48dfbcd91b1bb2ead4b44
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1929
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Change-Id: I88afad6d071e8b577c23da9c60392c60b3726514
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1895
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Non-KDEBUG kernels and kernel add-ons use atomic operations to acquire
and release the locks inline, so non-KDEBUG kernels/addons are only
compatible with other non-KDEBUG kernels/addons.
Following this change, though, KDEBUG kernels/addons should be able
to run under non-KDEBUG kernels/addons, too, since they always call
into the actual kernel functions and do not inline anything of
consequence.
This makes ARM64 target compile more files. This patch is one of
series of patches to support new architecture, as fixes in many
places are required just to compile the code.
Signed-off-by: Jaroslaw Pelczar <jarek@jpelczar.com>
Change-Id: Ia060612733cd3a0fcb781fec449da164ed635b8e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1807
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
This adds one (private) VFS function, and checks in all usages of
the vnode->cache as a VMVnodeCache that it really is one. (Generic
usages, for the moment just the ReleaseRef() calls in vnode
destruction, are intentionally not touched.)
This will be used by ramfs to set the cache from its own,
so that map_file() calls on a ramfs can work.
The patched errata are only the AMD ones FreeBSD patches
(it seems there are no Intel errata that can be patched
this way, they are all in microcode updates ... or can't
be patched in the CPU at all.)
This also seems to be roughly the point in the boot that
FreeBSD patches these, too, despite how "critical" some
of them seem.
Change-Id: I9065f8d025332418a21c2cdf39afd7d29405edcc
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1740
Reviewed-by: Jessica Hamilton <jessica.l.hamilton@gmail.com>
Sparcv9 runs Openboot in 64 bit mode, which means the cell size is
64bit. Use intptr_t where appropriate to make the open firmware calls
work.
Beware, some values are still 32bit, this matters for example for
of_getprop, if you get 32bits into a 64bit variables it will be in the
MSB of it (big endian only weakness...) and confuse things. See for
example in console.cpp, where the input and output handles are retrieved
as 32bit values. It seems wise to check the expected size when using
of_getprop in these cases, instead of just checking for errors.
Change-Id: Ie72ebc4afe7c6d7602a47478f0bfb6b8247004b8
Reviewed-on: https://review.haiku-os.org/c/1369
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Before this commit, *all* ConditionVariable operations (yes, all;
even Wait, Notify, etc.) went through a single spinlock, that also
protected the sConditionVariableHash. This obviously does not scale
so well with core count, to say the least!
With this commit, we add spinlocks to each Variable and Entry.
This makes locking somewhat more complicated (and nuanced; see
inline comment), but the trade-off seems completely worth it:
(compile HaikuDepot in VMware, 2 cores)
before
real 1m20.219s
user 1m5.619s
sys 0m40.724s
after
real 1m12.667s
user 0m57.684s
sys 0m37.251s
The more cores there are, the more of an optimization this will
likely prove to be. But 10%-across-the-board is not bad to say
the least.
Change-Id: I1e40a997fff58a79e987d7cdcafa8f7358e1115a
Even on 64bit CPUs it's a 32bit register.
Change-Id: I9a4de6eec225de19a90d70fae1382b662e530629
Reviewed-on: https://review.haiku-os.org/c/1625
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
B_OS_NAME_LENGTH is 32, char* is 8 (on x64), and this structure
has quite a lot of pointers in it so it is not like we really
needed to save those 24 bytes. Hitting malloc() in here is not
so great, especially because we usually have B_DONT_LOCK_KERNEL_SPACE
turned on, so just inline and avoid it.
Change-Id: I5c94955324cfda08972895826b61748c3b69096a
This was (following the packagefs changes) the number-one (by call
count) consumer of malloc() during the boot -- 52866 calls, and 100%
of them either 1024 or 1025 bytes!
Virtually all of these are ephemeral (indeed, the object_cache
stats after a boot with this patch shows there is only a single slab
of 64 buffers allocated, and most of them unused), so this is
probably a significant performance boost.
Change-Id: I659f5707510cbfeafa735d35eea7b92732ead666
If the buildbots were working, I would have been informed of this
about an hour after I committed it last night. But it seems they aren't.
Maybe kallisti5 will have some more incentive to work on that?
Cleans up some lock/get/unlock sequences, and makes it possible
for external consumers to get team structs (which will be necessary
for permissions checks.)
* Now matches the rest of the architectures.
Change-Id: I6699e0c8f729923770f136f2c9599185a685336a
Reviewed-on: https://review.haiku-os.org/c/1527
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Reviewed-by: Axel Dörfler <axeld@pinc-software.de>
* mutex_destroy() only checked wether or not there were waiters,
not if the lock itself was presently held by another thread.
Now we do, which should make #15015 panic much earlier instead
of trying to use freed memory.
* mutex_transfer_lock() and recursive_lock_transfer_lock() did
not check that the calling thread actually owned the lock.
Now it does, which should trigger asserts if anyone tries
to do this.
Copied from PPC with the hooks for Apple hardware removed.
To be completed with the actual PCI bus implementation for Sun machines.
This is where we start doing machine specific stuff, apparently.
Change-Id: I06af4de9621e9d40593d153642478d928083e49a
Reviewed-on: https://review.haiku-os.org/c/1364
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
* Kernel is 64 bit, and we won't need a 32bit load base.
Change-Id: I729bab01c8f71083002db061e153b0e5052b9a1c
Reviewed-on: https://review.haiku-os.org/c/1326
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Add a platform cleanup hook before starting the kernel. The openfirmware
and PXE loaders clean up their network stack there, while the other
loaders currently do nothing.
This closes ticket #6166
Change-Id: I34765892dfd9b2310c6af97c9ff7d414afae49e5
Reviewed-on: https://review.haiku-os.org/c/50
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Add empty implementation of timer, elf, vm, debugger support, to let the
kernel link.
Also add the kernel linker script.
Change-Id: If0795fa6554aea3df1ee544c25cc4832634ffd78
Reviewed-on: https://review.haiku-os.org/c/1108
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Previous commit adding these was merged very quickly, so here's one
more...
Change-Id: I23c424db7631db1f0ec48e2d0ae47c8409ae6af2
Reviewed-on: https://review.haiku-os.org/c/1088
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Gets the stage0 bootstrap to run.
Imlementation is probably nonsense at this point.
Change-Id: I10876efbb54314b864c0ad951152757cdb2fd366
Reviewed-on: https://review.haiku-os.org/c/1061
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
There was no synchronization of the check of the done flag and the
waiting thread suspending to wait for it. It was therefore possible that
the new team both set the flag and triggered the wakeup of the waiting
thread in that time window, causing it to miss both the set flag and the
thread resumption.
Use a condition variable instead.
Fixes#13081.
Change-Id: I93c45db8dd773fe42b45c4b67153bcd39e200d3b
Reviewed-on: https://review.haiku-os.org/803
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This file is included, directly or indirectly, by most of the
kernel-space C++ code, and so importing the entirety of "std"
seriously pollutes the global namespace.
So instead, just import "std::nothrow", which is the only thing
we really want in the global namespace. Tested on both GCC2
and GCC7 and seems to work just fine.
While I'm here, also update the include guards and copyright
header to match the standard format used elsewhere.
* Consolidate all fdt code into fdt bus_manager
* Build boot and kernel static libraries
Change-Id: I2a69cd7e1f1276999a80734ff12918fd49b599e5
Reviewed-on: https://review.haiku-os.org/440
Reviewed-by: Alexander von Gluck IV <kallisti5@unixzen.com>
memset uses rep stosb on x86 during boot, with memory
not set to write-combining, which makes it slow.
Instead we do aligned writes of 2 x four bytes at once.
Only clear the minimum of size and width * height * 4
UEFI framebuffer size can be huge, upto 512MB here,
and rep stosb seems to be around 25-30MB/s
This is written as generic as possible to work on
old compilers and different platforms, without
expecting boot memset to be optimized.
This makes it almost unnoticable compared to not
clearing.
It seems not all of the kernel includes this, but some use new/delete
anyway. Further, operator delete[] was not implemented at all.
Possibly fixes the ARM build.
Writes to videomem is slow without memory remapping
Can't do the mapping without leaving UEFI, so skipping
the clear. Afaict it should always be cleared by UEFI
This saves ~10 seconds of booting on my machine
(1920*1080*4 bytes)
EFI video mode (should have been it's own commit)
* Only do strcmp if there are enough params
* break when found
This reverts commit c558f9c8fe.
This reverts commit 44f24718b1.
This reverts commit a69cb33030.
This reverts commit 951182620e.
There have been multiple reports that these changes break mounting NTFS partitions
(on all systems, see #14204), and shutting down (on certain systems, see #12405.)
Until they can be fixed, they are being backed out.
* define compat_thread_info, compat_rlim_t, compat_rlimit and
compat_thread_creation_attributes to be used when applicable in compatibility
mode.
* handle 32-bit types in _user_spawn_thread(), _user_get_thread_info(),
_user_get_next_thread_info(), _user_getrlimit(), _user_setrlimit(),
other syscalls are compatible as is.
* init TLS for compatibility mode threads.
Change-Id: I483ba95e6198ddac9d240671bcb56fcd2ad831d2
* in load_image_internal(), elf32_load_user_image checks whether the binary
format requires the compatibility mode.
* we then set up the flag THREAD_FLAGS_COMPAT_MODE and the address space size.
* the compatibility mode runtime_loader is hardcoded with x86/runtime_loader.
* if needed, the 64-bit flat_args structure is converted in-place to its 32-bit
layout.
* a 32-bit flat_args isn't handled yet (a 32-bit team execs a 64-bit binary).
Change-Id: Ia6a066bde8d1774d85de29b48dc500e27ae9668f
* define compat_area_info to be used when applicable in
compatibility mode.
* handle 32-bit types in _user_reserve_address_range(), _user_get_area_info(),
_user_get_next_area_info(), _user_transfer_area(), _user_clone_area(),
_user_create_area(), _user_map_file(), other syscalls are compatible as is.
* _get_next_area_info() doesn't work well with a 32-bit address cookie (address
could be in 64-bit range). Instead use _compat_get_next_area_info() which uses
the area id as cookie, though the areas are not ordered by address any more.
Change-Id: Ic7519ca8824aa2d534b0f03ea75a1bf6ae321535
* handle 32-bit types in _user_send_signal(), _user_sigaction(), _user_sigwait(),
_user_set_signal_stack(), _user_restore_signal_frame(), other syscalls are
compatible as is.
Change-Id: I4c8dc47bfa80f36e363d444d2a5a7be6c621606d
* define compat_image_info, compat_extended_image_info
to be used for respective 32-bit types of syscalls in compatibility mode.
* handle 32-bit types in _user_register_image, _user_get_image_info,
_user_get_next_image_info, other syscalls are compatible as is.
Change-Id: Ibbd33e6796208dfa70d869e36bf745bc3e18d330
* define compat_flock, compat_timespec, compat_stat, compat_attr_info,
compat_fs_info, compat_fd_info to be used for respective 32-bit types
of syscalls in compatibility mode.
* handle 32-bit types in common_fcntl(), _user_read_stat(), _user_stat_attr(),
_user_read_index_stat, _user_read_fs_info, _user_write_fs_info,
_user_get_next_fd_info, other syscalls are compatible as is.
Change-Id: I5b372169fe142f67b81fd6c27e0627d5119ba687
* File locks created by flock should only apply for the file descriptor
that was used to lock the file. Another fd on the same file should then
be denied access (calling flock should fail).
* fcntl based locks, however, are in a separate namespace and are global
to a team.
* This issue was found when running webkitpy test suite, and should close
ticket #13795.
* Don't use session or team as comparison in release_advisory_lock(), as
that information might not be available anymore (e.g. when called from
Team::~Team()). This fixes#14121.
Change-Id: I9efb96cfcefe7e72b0060220c635a665e7e643cc
Co-authored-by: Axel Dörfler <axeld@pinc-software.de>
* x86 uses a commpage with 32-bit addresses, incompatible with the one used for
x86_64. For this reason, a compatibility commpage is needed to support a 32-bit
userland on x86_64.
* define ADDRESS_TYPE as a macro for addr_t (default) or uint32 (for the 32-bit
commpage).
* team_create_thread_start_internal() will use clone_commpage_area() with
KERNEL_USER_DATA_BASE or clone_commpage_compat_area() with
KERNEL_USER32_DATA_BASE, to setup the correct commpage.
* real_time_clock (in compatibility mode) also updates the compatibility
commpage with real time data.
Change-Id: I61605077ce0beabab4439ef54edd1eae26f26fd2
* define ELF32_COMPAT to enable ELF32 macros.
* add a flag ELF_LOAD_USER_IMAGE_TEST_EXECUTABLE to only check the format.
It will be used by load_image_internal() to check which mode to use when
loading an image.
* in arch_elf_relocate_rel(), switch to elf_addr instead of addr_t, which
would be the wrong size for elf32 on x86_64.
* the ELF compat loader reuses the relevant parts of elf.cpp and arch_elf.cpp,
excluding for instance load_kernel_add_on() or dump functions.
Change-Id: Ifa47334e5adefd45405a823a3accbd12eee5b116
* also adjust BOOT_GDT_SEGMENT_COUNT for x86, the definition is used by the
boot loader.
* add some 32-bit definitions.
* add a UserTLSDescriptor class, this will be used by 32-bit threads.
Change-Id: I5b1d978969a1ce97091a16c9ec2ad7c0ca831656
It was limited to a uint32 and could for example be overflown by the
slab MemoryManager that uses size_t on a 64 bit system.
This aligns the signature with create_area() that already uses size_t
for the size argument.
Note that the function is currently private, so the impact should be
limited.
SMAP will generated page faults when the kernel tries to access user pages unless overriden.
If SMAP is enabled, the override instructions are written where needed in memory with
binary "altcodepatches".
Support is enabled by default, might be disabled per safemode setting.
Change-Id: Ife26cd765056aeaf65b2ffa3cadd0dcf4e273a96
The scheduler uses the load tracking logic to compute the load of
threads to be enqueued into the run queue. The time delta between the
last enqueue and the next enqueue may grow very large for threads
that mostly wait on conditions. In such cases the int "n" period count
variable would become too small and wrap around, leading to an
assertion failure.
For this to happen, the thread in question would have to have slept for
at least ~25 days and then wake up. Threads often affected would be ones
waiting for some other process to end, for example shell threads waiting
for a long running process to exit.
Fixes#13558.
The base VMCache class changed to the generic_ types with their
introduction in in *2011* (435c43f591),
but these classes were never properly adapted. These functions should not
be called here (they panic() -- but the base class only returns B_ERROR,
so that is a difference at least.)
Found by Clang's -Woverloaded-virtual.
The UDP service does not own the UDP sockets. When shutting down,
inform the bound sockets that the service is no longer available.
This allows subsequent method calls to error out cleanly.
Signed-off-by: Augustin Cavalier <waddlesplash@gmail.com>
Add a cleanup function net_stack_cleanup() that calls a new NetStack::ShutDown() method.
Make sure this method works even if the network stack was never initialized.
Signed-off-by: Augustin Cavalier <waddlesplash@gmail.com>
