Before 2019, the entire ConditionVariable system was "giant"-locked:
that is, there was a single global lock that all ConditionVariable
and ConditionVariableEntry operations had to pass through. This of
course was not very performant on multicore systems and when
ConditionVariables see significant use, so I reworked it then to have
more granular locking.
Those patches took a number of attempts to get right, as having two
objects in separate threads that can each access the other not turn
into a deadlock or use-after-free is not easy to say the least,
and the ultimate solution I came up with erased most of the performance
gains I initially saw on the first (partially broken) patchsets.
So I have wanted to revisit this and see if there was a better way
even since then. Recently there have been a few reports of
ConditionVariable-related panics (apparently double unlocks),
notably #16894, and so that was reason enough to actually revisit
this code and see if a better solution could be found.
Well, I think I have come up with one: after this commit, Entries
no longer have their own lock, and instead accesses to Entry members
are almost always atomic; and there is now a case where we spin inside
Variable::_NotifyLocked as well as one in Entry::_RemoveFromVariable.
This leads to somewhat simpler code (no more lock/unlock dance in Notify),
though it is significantly more difficult to understand the nuances of it,
so I have left a sizable number of comments explaining the intricacies
of the new logic.
Note: I initially tried 1000 for "tries", but on a few instances I did see
the panic hit, strangely. I don't think the code that is waited on can
be reasonably reduced any further, so I have just increased the limit to
10000 (which is still well below what spinlocks use.) Hopefully this suffices.
Quick benchmark, x86, compiling HaikuDepot and the mime_db in VMware, 2 cores:
before:
real 0m23.627s
user 0m25.152s
sys 0m7.319s
after:
real 0m23.962s
user 0m25.229s
sys 0m7.330s
Though I occasionally I saw sys times as low as 7.171s, so this seems
to be at least not a regression if not a definitive improvement.
Change-Id: Id042947976885cd5c1433cc4290bdf41b01ed10e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4727
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
GCC still assumes that the dirent has no data past the end for some
scenarios here and still mis-optimizes things. Therefore, drop the
usages of unions altogether, and instead use a casted character array.
Additionally, use B_FILE_NAME_LENGTH for the array, not B_PATH_NAME_LENGTH,
and make sure to add 1 for the NULL terminator.
The lock entry is the first thing in the struct, so this is a no-op
change, but it is safer to do in case of changes, of course.
Spinlocks have been structures for quite a long time, so this was
probably just missed in the conversion.
we don't sample if the last sample is too recent and use the cached result.
Change-Id: I17ed29bda7fe7276f1a4148b3e1985c9d32ae032
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4101
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
This file should ideally contain only those things needed
across all system headers, even POSIX ones, and all other
declarations (B_* ones especially) should go in SupportDefs.h.
However, as nothing but riscv64 uses this right now, I've just
moved it to there.
GCC 11 treats [1] as a fixed-length array and not a flexible-length
array, and so some things that used direct strcmp("..", ent->d_name),
for instance, would be optimized out as being always unequal,
which was the cause of #17389. Using a real FLA informs GCC that
there is going to be more than one byte of data, and thus this
fixes that bug.
BeOS used [1] and not [0], possibly because it had to deal with
compilers (MetroWerks? Early GCC2?) that did not support FLAs.
GCC 2.95 does, using [0], and GCC 4 does, using [], so we can go
with that here.
(I did try using [0] for both, which seems to be OK with GCC 11,
but GCC 8 throws errors when d_name is dereferenced directly
as being-out-of-bounds. So, we have to use the #if here and give
newer GCC the [] syntax and not [0] to avoid that problem.)
The real question probably is whether or not we should backport
some variant of these changes to R1/beta3, as software at HaikuPorts
very well may run in to the same issue. (The alternative workaround
is to compile with -O1 and not -O2 for any affected software.) But
maybe this is an argument for keeping with the beta4 schedule of
this coming January...
At present, it does, but that is an oddity we have preserved from BeOS
that the next commit is going to remove. (This commit thus wastes 1 byte
without the following one.)
Most changes are pretty straightforward: only a +1 is needed,
and a few removed from sizing calculations. Some filesystems like UDF
originally passed back the length with the \0 included, so they have
been adjusted further. UFS2 had some other sizing problems which are also
corrected in this commit.
Our dirent structure is "slim": it has a flexible-length array at the
end which must be allocated to whatever size the consumer wants. However,
we use [1] there and not [0] or [], which meant GCC thought it was not
a flexible-length array, and so it optimized various string accesses
that it assumed must be always false. Among these was BDirectory's
check for "." and "..", and so that resulted in infinite loops.
When changing our dirent structure to a proper FLA instead of [1],
GCC then throws errors on LongDirEntry as it has data "after" the
FLA; which is what we want, but there is no way to tell GCC that.
So now we use a union instead, which is the proper way to statically
allocate a FLA.
This is part of #17389, but the real fix requires changing our dirent
structure, which is coming in a separate commit.
As suggested by PulkoMandy. This was done before my commits yesterday,
but those just reverted patches that had only been in since May,
so it's not clear how much this is actually needed. Nonetheless
it seems like the more correct thing to do.
While FreeBSD and glibc feature-guard it, they also feature-guard
a lot of other things that we don't, and musl does not guard it,
so it seems more than safe enough to leave it unguarded.
Fixes compilation errors with GCC 11. (The other possible solution
was including features.h in more places, but this seems simpler.)
* A few things need alignment, instead of forcing them all
to align themselves, support alignment of the kernel_args
* Default of 1 is "no alignment"
Change-Id: Iff05dcec8adaa963c8444d701464ea11616062f6
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4698
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
* Catch errors and report them in bus parsing code
* Align the FDT kernel_arg to 8-bytes
* we still choose BSD-2-clause :-)
Change-Id: If2a88b7f131025ff1c1a2d903ed52f039e5bbcb5
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4694
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
In most cases we don't need to use the complete display_mode struct and
we just need the timings. This will avoid future confusion between the
virtual width/height and the actual display timings, if we implement
scrolling someday.
Change-Id: I6c4430b84130b956a47ea0a01afb0843f5a34fd2
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4665
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This way it becomes much easier to write multiple console implementations
in one bootloader.
Tested for bios_ia32 and efi.
Change-Id: I67134f5c3de109b15d46898864ba7f51c6592afc
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4642
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
* Delete dropped out networks.
* Add in newly discovered networks.
* Add static (aka class) compare method to WirelessNetworkMenuItem
that is used to sort items by signal strength descending.
Add == operator to wireless_network struct to determine if
existing items have a known network attached.
Remove the non-network items from the menu, save them, sort
network menu items, then add non-network items back into the
menu.
Update NetworkStatus preflet to use same compare method as Network
preflet. signal_strength_compare function had a bool return value
instead of int which worked to sort items the first time, but does
not work on successive compares.
By not deleting and recreating the menu items each Pulse(),
the Network preflet no longer crashes on update. The menu flashes
on update still but doesn't crash.
Fixes#12024
Change-Id: Ie5b22cea4e66350b9c5df8e3b8de266ede50ad6d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4243
Reviewed-by: John Scipione <jscipione@gmail.com>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Reviewed-by: Axel Dörfler <axeld@pinc-software.de>
... methods which call the respective methods in BList.
These convinience methods allow you to sort a menu of menu items
via a compare function, swap two menu items, or move a menu item
to a new index. Update items layout if menu is open.
Previously there was no easy way to rearrange menu items in a menu.
Change-Id: Ice3d6e5404e895196d8bd32d696dce7c55bd72d4
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4296
Reviewed-by: John Scipione <jscipione@gmail.com>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
This cuts out almost 40,000 lines of these headers. (I did something similar
in the atheroswifi AR93xx/94xx driver when importing it from FreeBSD,
which had a lot more than 40,000 lines.)
The code in this module was derived from the one in driver/tty. However,
the driver uses a shared lock between the master and slave side of a
TTY, and this was changed to use two separate locks. The approach with
two locks does not work. It seems the change was unfinished and the
second TTY was never locked. But attempting to lock it will result in
lock inversion problems, unless we do complicated things (try to find
which of the two TTY is the master side, and lock that first, for
example). It is simpler to restore the shared lock as used in the
driver.
To set up the shared lock, I modified the tty_create function to take a
pointer to the master TTY when creating the slave. Maybe it makes more
sense to create both sides in the same call, create_tty_pair?
However, this does not work as easily as I wanted, because there is some
recursion going on: at least in one case, the tty_control function is
calling the driver's tty_service function, which in turns attempts to
call back into tty_control for the "other side" TTY. To handle this
case, replace the mutex with a recursive_lock.
Fixes#17091, where the root problem was access to
other_tty->select_pool without locking. This was also made unconvenient
to debug because select_pool objects are self-deleting, when the last
item in the pool is removed. As a result, the code accessing it without
log would suddenly find out that the data it was accessing had been
freed and erased.
This also makes the TTY code in driver/tty and generic/tty a bit more
similar than it was before, and brings us one step closer to merging the
two together. There are still two main differences and I don't know
enough about TTY to decide if they are important, and which version
should be kept:
- The driver has extra code for "background" read and write. I don't
know what this is used for.
- The driver has a single "settings" instance shared by a master and
slave TTY, while the module has two separate instances, but seems to
copy one to the other. I'm not sure which approach is correct.
Change-Id: Ie2daddd027859ce32ba395af76b4f109f8b984b0
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4604
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
These will be needed to implement custom modes in the VESA driver.
Change-Id: I9b52de691baa14e1f1a3ccce500ced9bb040b113
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4622
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Now that it is not used anywhere in the source tree following
previous commits.
Change-Id: Id2fc417a0658d09148e99587c613a928f1fbe4c2
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4611
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
This reverts commit 3e8376c6dd.
Reason for revert: Bootloader currently fails to load kernel
It should be added back once the kernel can start.
Change-Id: Iebefbf8681aff4dff09cef7b7eb832b61f7789c7
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4579
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Someone on the internet found out gcc only understand posix_memalign.
The alloc_align attribute may be applied to a function that returns
a pointer and takes at least one argument of an integer or enumerated
type. It indicates that the returned pointer is aligned on a boundary
given by the function argument at position.
Change-Id: I4b0af6ef3020da1fb460652117286193d5d72f1e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4514
Reviewed-by: Fredrik Holmqvist <fredrik.holmqvist@gmail.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Change-Id: Ifadd47204be1ec688017a567d43dca38c80bd1df
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4431
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: Fredrik Holmqvist <fredrik.holmqvist@gmail.com>
Set first stack frame return address to
<commpage>commpage_thread_exit, so it will be called
when thread entry point returns.
Change-Id: Ide5cde8d4501eb7241e03ff4052174e984e78870
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4493
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
These were used in function_remapper.cpp but can be used elsewhere too,
so move them to a private header. Also use them for the stack protector
hidden function definition (probably not so useful since gcc2 doesn't
support using the stack protector anyway?).
The gcc2 way to make a symbol hidden is to manually generate the .hidden
directive in the assembler output. This is not perfect: it is hard to
use for C++ functions and methods (manual mangling of the name is
needed), and inline assembler can only be inserted inside functions. But
the alternative is patching gcc2 to add support for the function
attribute, and I don't want to dig into that today.
This is an old version of libc++ that was imported in an early attempt
of building Haiku with clang. It is currently not used for anything. In
fact there never was a Jamfile to build it.
* PCI ID's from Linux
* There are a bunch of NAVI quirks around 3d rendering pipelines
but not many around modesetting (which is the only thing we care
about)
Change-Id: If63e31fe1d37d9d95f2a71c222a4cda7a2914a5e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4467
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Only one code change: for some reason, GCC chokes on the cr3 functions
as macros (throwing errors about invalid registers.) The BSDs have them
as inline functions instead, so they are converted to that here.
Tested and working. There seems to be about a 10% decrease in CPU time
on some compilation benchmarks that I briefly tried.
Change-Id: I31666297394d7619f83fca6ff5f933ddd6f07420
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4515
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
these colorspaces are packed as RGB or RGBA, not BGR or BGRA.
RGB48_BIG and RGBA64 only differ in the endianess of the channel the 2-byte value.
this is a big difference with RGB24_BIG and RGBA32_BIG, in which case _BIG
means the order is RGB (BGR) and not BGR (BGRA).
BGR48, BGRA64 could indeed be added, if needed.
I chose 0x11 and 0x12 arbitrarily, but given the order of channels 0x1011
and 0x1012 might make more sense. This would mean using another bit for "real"
bigendian colorspaces.
Only the color conversion to 32-bits is implemented.
Tested with the RAWTranslator modified to output 16bpp with success.
Found some references in enum AVPixelFormat in libavutil/pixfmt.h.
Change-Id: I4b023dec85d01f1e63e1b053139e5bb5d263a0e0
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4468
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
It was removed in hrev55422 as we never had declared it in any headers.
But it seems some software came to depend on it anyway. Reinstate it,
and add a declaration for it, behind _GNU_SOURCE.
This is a source compatibility break from BeOS, but should not
be an ABI one (I checked, the symbols are identical.)
Also use "= {}" in the definitions of the fields. We use this
in plenty of places in the kernel, so it should be OK for GCC2.
Change-Id: Ibe05b2236d46024d7b4563ae16e1cc7140fed965
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4434
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This removes the use of the destructor in the move assignment operator, as it
may rely on undefined behaviour from the compiler. Additionally, some duplicate
logic to dereference and free a shared string has been unified under
_ReleasePrivateData().
Change-Id: Ie9f51d598c734f83cd0fba49b651315c6e9c8aac
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4440
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
This implements the "rule of 5" for this type. While the copy operation for
BString was already using shallow copies of the underlying data, this change
further optimizes moving the data from one object to another.
While it is not the intention to implement move semantics to all types in the
legacy Haiku/Be kits, data types like BString are good candidates, because move
operations are often useful when working with data within an application.
In this implementation, the internal data of the string object will be set to
NULL, thus leaving an empty string.
Change-Id: I16bf9424f9b17f622b0b57659b80628e18760288
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4428
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
* Drop ArchUART8260 layer to reduce complexity. It's whole
existance in life was to adjust the mmio alignment.
* Fold architecture mmio alignment into DebugUart
* We could potentially pass a Init(int mmioAlignment)
arg in the future if the macros get too messy.
* Move Barrier code back a layer into DebugUART
* Fixes the arm uart and EFI build
Change-Id: I0f127d902993e9f6e6a03cac8c7c37c0363134bf
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4422
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
The USB Kit uses it, so this allows the USB Kit to stop including
USB3.h.
Change-Id: Ifde025ec41bef92013fda0440d60b7216cfdbe4a
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4413
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
* We really should get out of the habbit of making up
our own architecture defines.
* __riscv with an additional __riscv_xlen is the
standard that developed... let's just roll with it.
Change-Id: Ieb777d48340ae25a6d66f66133afa0ec5c6da9b6
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4402
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Including thread.h brings a massive array of things with it from
the kernel thread arch headers, team and thread definitions,
hash tables, linked lists, Referenceable, etc. that the vast majority
of AutoLock.h consumers neither want nor need.
So, put these in a separate header, and adjust all consumers of these
lockers to include the new file.
This change exposes the fact that a lot of files were inadvertently
making use of headers included indirectly through thread.h. Those
will be fixed in the next commit.
Previously these were just using the raw function name, which led
to markers like "Slab_begin". Now we prefix RANGE_MARKER_ so there
is absolutely no chance of confusion, and the symbols are clearly
visible in dumps.
Also add a note that the kernel must be built with -fno-toplevel-reorder
for these to work. (It seems when this was implemented, GCC had not yet
implemented top-level reordering.)
They are only used for debugging with the tracing system in a handful
of places, and -ftoplevel-reorder is enabled with optimizations for
a reason, so it makes more sense just to note this and not to enable
that option by default (i.e. in the off chance someone will want to
use these in non-debug builds, like I did.)
If the timeout is already >= B_INFINITE_TIMEOUT, we do not need
to do any of the following math (which would usually overflow anyway)
and can leave the timeout alone.
Spotted by kernel undefined behavior sanitizer.
Define thumbnail attributes in Attributes.h:
Media:Thumbnail to store the thumbnail,
Media:Thumbnail:CreationTime to see if thumbs need to be regenerated.
Store 128x128 thumbnail in attribute, for icon sizes smaller than
128x128 down-scale the 128x128 thumbnail. Use B_FILTER_BITMAP_BILINEAR
to down-scale the image using the bilinear scaling algorithm which
creates nicer looking thumbnails than the default scaling algorithm.
Store thumbnails as WebP images which compress smaller than PNGs and
fit in the inode better at 128x128.
Check the file's modification time in GetFileIconFromAttr() and compare
it to the thumbnail creation time. If the file has not been modified
since the last time we generated thumbnails return the thumbnail from
the attribute, otherwise fetch a new thumbnail with GetThumbnailIcon().
Add "Generate image thumbnails" Tracker setting. Default is turned off
for now. To generate image thumbnails you must first turn this setting
on in Tracker Windows preferences.
Spawn a get_thumbnail() thread to generate thumbnails and retrieve them
later on from the window thread to fill out into the icon. This should
improve responsiveness of generating thumbnails from a folder with a
lot of images. The generator thread will write the thumbnail data to an
attribute if on writable BFS volume.
If not on writable BFS volume, the generator thread will send the data
back to the original thread through a port by calling write_port().
When the thread is finished creating the thumbnail it sends a message
back to the Tracker application thread to update the pose which
instructs the window thread to look for an thumbnail. It either finds a
thumbnail in an attribute, or picks up the thumbnail data that has been
sent through write_port() using read_port().
This works on both read-write and read-only BFS volumes but it still
depends on the presence of a BEOS:TYPE parameter to have been written
to the volume before it became read-only. Thumbnail generation does not
work on other read-only volumes for example an ISO-9660 CD, but it does
work on read-only BFS volumes for example the BeOS R5 CD.
Move BPrivate::CheckNodeIconHintPrivate() from BNodeInfo to Tracker
Model CheckNodeIconHint(). Create Model::CheckAppIconHint() and look
for a vector icon or mini and large icon in that method. Check that
the base type is directory, volume, trash, desktop, or if executable
call CheckAppIconHint().
Add 1 to temp_name to fix the following warning:
src/kits/tracker/FSUtils.cpp:2437:12: note: 'snprintf' output 3 or more
bytes (assuming 267) into a destination of size 266
Rename temp_name to tempName following our style guidelines. Use
strlcpy() and strlcat() instead of strcpy() to safely copy the string.
This fixes thumbnail generation on 64-bit Haiku.
Change-Id: I7f927a5a1f8cf65e4b1aa1e0eb55bbfae87fd969
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3163
Reviewed-by: John Scipione <jscipione@gmail.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
* This models the CpuInfo into a cross-architecture
platform_cpu_info
* Originally I was looking at merging this with "arch_cpu_info"
however that is "overall cpu" while CpuInfo is "indivial core
information" packed into an array.
* Since every dtb platform will report individual cores in fdt,
having a common cpu core info struct with at minimum the core
id makes sense.
* This could likely be refined further to some kind of core info
packed inside of arch_cpu_info, but this will fix arm,arm64,etc
for now until someone wants to dive into that.
Change-Id: Ia18a352403cd0da7130c1e637fc205d4311478ef
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4363
Reviewed-by: Fredrik Holmqvist <fredrik.holmqvist@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Change-Id: I6cb31760519c8ba4542d217d6e68439602eda558
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4356
Reviewed-by: Jessica Hamilton <jessica.l.hamilton@gmail.com>
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Change-Id: I4b8f69271ede117701725f9cce30de5bb8ba30bb
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4332
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: Jessica Hamilton <jessica.l.hamilton@gmail.com>
It allows to call destructor function stored in struct object such as
device_manager_info::put_node.
Change-Id: If9162f2f449d2b1c52c39509fa8732f21debf04a
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3484
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
- Remove Pause/Resume functions. They are not possible to implement (the
server would time out)
- Fix SetContext(NULL) to do the right thing.
Change-Id: I25ba09bb01ea0fe8a85d774611b33be7dc192028
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4245
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Niels Sascha Reedijk <niels.reedijk@gmail.com>
The APIs for this were introduced in ICU 63, so we'll need an update.
ICU 63 does not build with gcc2, so this method is disabled there.
Change-Id: Iabe49509ed6d4e578560d497d3ca336a97db4625
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1874
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Fixes:
* scsi: Fix a bug that caused the device capacity to be set
to an undefined value for some large SCSI devices when
READ CAPACITY (16) was used
* ahci: Fix VPD page reporting so that it does not return
undefined values
* ahci: Set the write bit to true when sending a DATA SET
MANAGEMENT (trim) command to a device. The command would
otherwise fail and time out on some devices.
Improvements:
* scsi: Extend the READ CAPACITY (16) support to also
include logical block provisioning information
* scsi: Prefer READ CAPACITY (16) over READ CAPACITY (10)
on devices that are expected to support this command
* scsi, ahci: Enable trim on SCSI and SATA devices that
are expected to support trim and which correctly report
trim support
* ahci: Redo the implementation of the SCSI UNMAP command
* scsi: Redo UNMAP-related code
* scsi: Add support for UNMAP via WRITE SAME (10) and
WRITE SAME (16) commands
* When copying trim ranges between different data types,
make sure that the values don't change (detect overflows)
* Report the number of trimmed blocks even if the trim
operation fails
Change-Id: Ie5fc993bbbc19546b4308138ba10184bf7b9986a
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4157
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Fixes:
* Use uint64 instead of off_t when handling offset and size
of the trimmed range in the fs_trim_data structure
* BlockAllocator::Trim: Correct the size of a buffer
* ram_disk, mmc: Do not trim past device capacity
Improvements:
* BlockAllocator::Trim: Because the received offset and size
are ignored by BFS (the functionality is not implemented yet),
return B_UNSUPPORTED if the range does not cover the whole
partition
* ram_disk, mmc: More accurate calculation of the number
of trimmed bytes
* devfs: Add a uint64 version of translate_partition_access()
Change-Id: I24f4c08674f123ad33a5fef6e28996a4ada6ff0d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4155
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
* Makes the case where the loader and the install differ by
release type, so that the icons are rendered in the same
position
Change-Id: I01e48109ce127b202ce5e05544aa2d5a495ed53e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4162
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
This also keeps the functionality of hrev53848, which simplifies the
list of disks searched for bootable partitions; however, it maintains
the previous behaviour of platform_get_boot_partitions that continues
to iterate over a list of possible boot partitions, which should
allow finding a bootable BFS partition better in more circumstances.
Particularly, there are numerous reports of the UEFI loader entering
the boot menu despite it finding a bootable partition, which this
should address.
EFI's device_contains_partition is also structured such that it
compares the disk GPT table of the partition the loader is
querying of the EFI disk's GPT table, in the case that there are
multiple disks, as the most reliable method of comparison, with
a generic fallback for non-GPT disks, which will be less reliable.
This reverts commit 0d932a49ad.
Change-Id: I5fac8608035d56b8bb4dc6c3d495ec6db42fa9b7
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4149
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Theorically we could define __STDC_NO_THREADS__ internally in GCC, but
threads.h probably will be added in the future, and the compiler won't
notice.
Using this stdc-predef.h header and including it in GCC would solve this
nicely, and saves us from hardcoding.
The header can eventually be removed when obsolete.
Change-Id: I29aa58686e3c45449dc63e02e5a9e13a960b9090
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4097
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Don't resize text view in FrameResized() if resizable, this is done
in _AutoResize() instead. Set text rect width to width of max line when
word-wrap is off. Text rect width shrinks to the width of the text
matching behavior of BeOS R5 and previous Haiku. This fixes Tracker
Edit name.
Limit max width to column width in list mode or 30em in icon mode.
Filter paste messages limiting to max width in Tracker Edit name.
General BTextView fixes:
As a consequence of the text rect shrinking to fit the text, adjust
highlighting to go at least to edge of the view even if text rect width
is narrower. Extend the invalidation area beyond text rect when
redrawing to include highlighted areas.
Text views behave properly when overflow occurs i.e. when you type
text off the end of the text view. The text is nudged over as you
type/scroll so that the previous text is visible. This sorta worked
before but now works better.
Fix text rect centering by replacing switch with
BLayoutUtils::AlignOnRect().
Coalesce consecutive draw calls when inserting and deleting text to
prevent flashing for example when resizing the window. Redraw text
when the text view scrolls fixing a bug I noticed in StyledEdit.
Workaround negative height Beezer bug.
Fixes#16642, #16476
Change-Id: I2d32d6039944d2dc3218ce4de71f2966cc98c866
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3642
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
POSIX defines this structure but specifies only two fields (which we
already implement). However, both the *BSD and Linux have agreed on
some more fields, which are often assumed to be there by applications.
The benefice of having compatibility fields is
greater as having to patch every other software at HaikuPorts.
Change-Id: Ie28ca2e348aa16b4c57eb3498eb62175100d9b9d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4083
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Niels Sascha Reedijk <niels.reedijk@gmail.com>
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
* This will break previus ways to store settings (as it only stored a struct)
* Now we use BMessage to save data.
* Added some stuff to SettingsMessage.
* Fix a bug in BluetoothSettingsView::_GetClassForMenu() and SettingsMessage::SetValue
Change-Id: I6a0fa1564e78460258f480947592eb4007985007
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3887
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
the AC flag in eflags/rflags, pushed in the iframe by the CPU, is kept intact after handling the exception, since the fault handler is run with the faulted iframe and does a simple jump. The AC flag would otherwise be set until the syscall returns to userland.
Change-Id: I24f763032ab98029dd162fb411e1541586451606
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4040
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
It is not present in BeOS R5 and it just call unload_driver_settings.
Replace delete_driver_settings usages with unload_driver_settings.
Keep the symbol on x86 for binary compatibility.
Change-Id: I1382710e3a4cb5c65d1249ea0e5880891e6800e4
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3485
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
only available on gcc 4 and upper.
Change-Id: Ifc059d5fedd15a5aeda1514312fbbf98a72d3128
Reviewed-on: https://review.haiku-os.org/c/haiku/+/4005
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
It cause adding new entry to executable init array for each translation unit.
Change-Id: I1e2d7946da03c001de7721948bc9af8188e8b317
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3981
Reviewed-by: X512 <danger_mail@list.ru>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
* These are pretty much the 10G standards that have
any potential for usage on a desktop system.
Change-Id: I2cb49f41ca61e82e091d042f877ee2f1acb9c4ec
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3900
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
The ioctl call cannot know the expected number of arguments
because it depends on the specific ioctl being used, and the
same value could be used to do different things by different
devices. Without knowing that, it is not safe to use va_arg
(at best a random value will be read from the stack, at worst,
it will just crash).
This changes the implementation of ioctl in two ways:
- For C++ code: a 4 argument function with default values
for arguments 3 and 4.
- For C code: wrap arguments 3 and 4 in a struct with the
help of a macro, providing something that behaves like the
C++ version.
So, with this new code:
- Calling ioctl with only 3 arguments sets the 4th one to 0
- Calling ioctl with only 2 arguments sets the 3rd and 4th to 0
- Calling with 1 or 5+ arguments is a compile time error
The existing ioctl symbol is preserved for ABI compatibility.
Change-Id: I6d4d1d38fccd8cc9bd94203d3e11aeac6da8efc3
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3360
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
this is needed for GCC to build after the aligned_alloc introduction
Change-Id: Ieae32a050cc000561107c4a07cf10c912a196152
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3896
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
this adds kernel & libroot stack protector hooks. it uses /dev/random in userspace.
A configure option --enable-stack-protector is added to activate -fstack-protector
on selected system components (ATM apps, kits, servers).
Change-Id: If3a2920ba9aa0a85eaff4ba6778947f8c76ade31
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3895
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
* Thanks to kottan I notices something was of..
* Added uInt16 and uInt64, I was missing UInt16.
Change-Id: Id136dbb5a81392a7a694ac1fbbd9aefbd7f77af3
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3888
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
should help with bug #16929, untested
Change-Id: Ia7b9b6cc8e84e2377d79c0edd1c278cdf74d869b
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3891
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
* Move SettingsMessage
* Remove SettingsMessage from MediaPlayer and WebPositive
* Use the central SettingsMessage in MediaPlayer and WebPositive (Later Bluetooth)
* Fix a Jam file.
Change-Id: I3bb82a40082c5ece5c2aea2468a77bcd9f15ce77
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3856
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
* Inactiveate button and things that don't do anything.
* Those button and controlles that hade any actions in the Bluetooth Pref now saves in settings.
* Fix some windows/views.
* Fix Copyright in last Bluetooth commit.
* Last commit before we move saving settings with BMessage.
* Changed PoupMenu to BOptionPopUp
Change-Id: I32b85f1985b558d24b294a184665e08e6ce18a7d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3829
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
This is not my code but from ticket #9265
* Made a picture of how it looks, old left andnew right. https://imagebin.ca/v/5wIe6TIMzw4C
* Think we have a bug somewhere and don't store the name of the Bluetooth device (shown i the image).
* I have made som small changes but other than that it's the same code as in the ticket
* Ran the src/tools/checkstyle/checkstyle.py to get som style stuff, probably missed some anyway.
Change-Id: Ifeb75c8ad890f541e100cdcf78b394675a48ada9
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3825
Reviewed-by: Fredrik Modéen <fredrik@modeen.se>
Various projects, both commercial and OSS, began to use inclusive
terminology. There is no reason to not do it.
In Haiku, bootloader uses Blacklist, which is recommended to replace
with Denylist or Blocklist. I think Blocklist is appropriate here,
since it's a list used to block offending driver at boot.
Some strings remain unchanged for compatibility with previous naming,
but this change prepares for later removal of these (once everyone has
updated their kernel and bootloader).
Change-Id: Id9105ff5e9fcb866000355089b5ef97bf63ee854
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3145
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
a protection_max attribute is added in VMArea.
a read-only opened file already can't be mapped shared read-write at the moment,
but can later be changed to read-write with mprotect() or set_area_protection().
When creating the VMArea, the actual maximum protection is stored in the area,
so that it can be checked when needed.
this fixes a VM TODO.
Change-Id: I33b144c192034eeb059f1dede5dbef5af947280d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3804
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
The standardized version of readv() and writev() take an int as the third
parameter. Arguably a size_t makes more sense, but the standardization bodies
decided otherwise.
The non-standard functions of readv_pos() and writev_pos() have been updated
for consistency. The corresponding _kern_readv() and _kern_writev() internal
functions continue to take the size_t parameter.
The ABI will not change, even though on 64 bit machines the size of the count
parameter will change from 8 to 4 bytes.
The actual use will be slightly different. Like with the size_t argument type,
it will not be possible to give a count lower than 0. If the value is less than
0, then the B_BAD_VALUE/EINVAL error will be set.
Change-Id: I949c8ed67dbc0b4e209768cbdee554c929fc242e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3770
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Do the final installation operations for all the packages in the
/system/packages directory when the OS is booted for the first time.
This will run their post-install scripts, create users, groups and generate
settings files (marked with a package version attribute). Previously we just
ran all the shell scripts found in the /system/boot/post-install directory
(don't do that as much now).
Fixes bug #14382
This patch has simpler code flow in CommitTransactionHandler::_ApplyChanges
Tested on 32 and 64 bit systems. Once it's official, need to remove the
open_ssh redundant post-install script that creates users etc. from HaikuPorts.
Now we can notice bugs like package version attributes on settings files aren't
fully working. :-)
Didn't remove special case for add_catalog_entry_attributes.sh since it
still does stuff that the build system doesn't do. Might be able to add
that script as part of the Haiku.hpkg. See change 3751 for removing it,
https://review.haiku-os.org/c/haiku/+/3751
Change-Id: I3807b78042fdb70e5a79eca2e2a45816ece0236f
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2342
Reviewed-by: Alexander G. M. Smith <agmsmith@ncf.ca>
Reviewed-by: Niels Sascha Reedijk <niels.reedijk@gmail.com>
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Updates BNumberFormat to be able to format
percentages. Also re-introduces some unit
tests and updates the BNumberFormat ones.
This doesn't actually fix#16312 as the
defaults for percentage formatting don't seem
to track the selected language, but goes part
way there.
Related to #16312
Change-Id: Id6ddf426ce5571f4e8513c0eb1663cf42ac53cb1
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3767
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Previously, BUrlRequest returns data received via a callback that can't
return any value. This approach have several issues:
- It's not possible to signify failures to the request.
- Users have to implement custom listeners just to handle the common
case of outputting to a buffer/file/etc.
- The received data has to be serialized into BMessage when
BUrlProtocolDispatchingListener is employed. This can cause a
noticible slowdown in real-world scenarios as evident by #10748.
With this change, BUrlRequest will output directly into a BDataIO, which
exposes a richer API for request handlers to work with (for example a
BitTorrent client can request a BPositionIO for non-linear data
delivery), as well as simplifying common cases for users.
The adaptation only requires one additional API:
BHttpRequest::SetStopOnError(). This API simply instructs the HTTP
request handler to cancel the request if an HTTP error is occurred.
Change-Id: I4160884d77bff0e7678e0a623e2587987704443a
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3084
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
The switch to make BUrlResult serializable was debuted in
f9e1854f19 with the rationale is that
BHttpRequest auto-redirection might cause the headers to become
obsolete by the time a client process the BMessage received from
BUrlProtocolDispatchingListener.
With the change to BHttpRequest to not notify listeners when
auto-redirection is enabled, this is no longer the case and the
serialization code can go away now. This simplifies BUrlResult and its
subclasses, and gain us some performance for clients using
BUrlProtocolDispatchingListener as the result object no longer has to be
serialized.
This also change the ABI of BUrlProtocolListener::HeadersReceived to no
longer passing a BUrlResult.
Additionally, BUrlResult and BHttpResult now express the size of the content
as an off_t, thus allowing results larger than 4 GB.
Change-Id: I9dd29a8b26fdd9aa8e5bbad8d1728084f136312d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3082
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
irssi uses the unichar typename with a different type.
Change-Id: I2ee9dcf685f4832dd1b3c7d68f4686d69de50fb7
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3656
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Reviewed-by: Fredrik Holmqvist <fredrik.holmqvist@gmail.com>
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Further removal of the use of custom list class;
this time with the package lists.
Relates To #15534
Change-Id: I1f01ed9d5ddbd7754097ce0adbf505d6ba17fd2f
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3732
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
- Unify storage of "FPU" registers between debugger and signal handler
to use xsave format on both sides
- Handle YMM registers in Debugger (they are the same as XMM, but wider)
Tested:
- The system still boots with and without AVX
- The hello_avx test program can be debugged and the full value of YMM is visible
This changes the API of vregs in signal.h but not the ABI (structure are
declared differently but memory layout is the same). This changes the
API and ABI of arch_debugger.h for x86_64, but I don't think anything
outside Haiku uses it (did we ever have a 64bit compatible gdb?)
Change-Id: If93680ffa0339c19bab517876b4e029f5d66b240
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3038
Reviewed-by: Rene Gollent <rene@gollent.com>
This change only applies in libnetservices.a. The implementation in
libbnetapi.so will use the original definitions.
Change-Id: I0aaa5a40af5fbcafaf233c32206cb4af862f8141
Reviewed-on: https://review.haiku-os.org/c/haiku/+/2465
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
The BGeolocation class uses the network interface in libnetservices.a, so it
is moved here for now.
This will break any out of tree projects that depend on it, but it is a source
incompatible change only.
Change-Id: I6f5b1332eb87ad37dd33fbe09fdb11b16f7f26e4
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3670
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
In order to prevent classes between libnetapi.so with the legacy API and
applications using the libnetservices.a library, the latter will have the
classes in a distinct namespace.
In the implementation, both libbnetapi.so and libnetservices.a will use the
same header and source files. If LIBNETAPI_DEPRECATED is defined during build,
the headers and source will have binary compatible behavior. Otherwise, the
classes and other objects will be put in the HaikuExt namespace.
In order to build the libbnetapi.so and libnetservices.a with the proper
build configuration, there is a stub `src/kits/net/libnetapi_deprecated` folder
that applies the special configuration to the source files.
Currently HaikuDepot, Webpositive, libshared.a and the http_streamer add on
use the compatible API in libbnetapi.so.
Change-Id: Ic73e9f271ef75749adda46f6f72e9a0b2851b461
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3667
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
These classes have been moved to the public API too soon, and they need some
more time to mature before they can be declared stable.
Change-Id: I9c52a8e6cc103922abde7a6b911fe0c3e6bf5700
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3665
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Fixes bfs part of #10336. Untested on SATA (don't have a testing drive
to sacrifice) but working fine on SD/MMC.
This requires moving the copy from kernel to userland into the devfs. As
a result the code in the disk drivers becomes a bit simpler.
Also add some documentation for the common ioctls to implement for a
disk device.
Change-Id: Ie84b6a1d293828d33902a64b3c9d4b19aa6eacb1
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3640
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Store the bus cookie in the mmc_disk driver and pass it to the bus
manager when executing commands. This avoids calling into the device
manager at each read and write operation. The code to get the cookie
from mmc_disk isn't so nice since it needs to access the grandparent
device (the mmc bus root), it would be simpler if this cookie would be
available directly from mmc bus devices.
We can get card removal and card insertion interrupt at the same time
due to insufficient hardware debouncing (the SDHCI spec says we
shouldn't, but it happens on Ricoh controllers. Can't blame them, they
don't advertise themselves as compliant with the spec). So, check the
card status from the interrupt handler and ignore the incorrect
interrupts.
Fix unreliable card initialization: power must be turned on before
starting up the SD clock. Remove a now unneeded delay that was added in
an attempt to avoid initial instability.
Change-Id: Ibd8d051da1a1d859f3924ee535f4a05d9b6398d4
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3639
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
It works, but performance is still unexpectedly low (getting about
50kB/s write speed) with almost no CPU load.
Change-Id: I7da3ee70c8b379c4e6c2250d67f880c78635874f
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3630
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
* otherwise the signal to be handled might be blocked. fixes#15193
* also remove automatic syscall restart on _kern_select, to match Linux and
BSDs behavior: this fixes parallel build with newer gnu make, which happens
to use pselect.
* also remove automatic syscall restart on _kern_poll.
from https://man7.org/linux/man-pages/man7/signal.7.html
"The following interfaces are never restarted after being
interrupted by a signal handler, regardless of the use of
SA_RESTART; they always fail with the error EINTR when
interrupted by a signal handler: ...
select(2), and pselect(2)."
from https://notes.shichao.io/unp/ch6/
"Berkeley-derived kernels never automatically restart select."
Change-Id: I3e9488f60c966b38d427f992f06e6e2217d4adc5
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3636
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Reviewed-by: Axel Dörfler <axeld@pinc-software.de>
* otherwise the signal to be handled might be blocked. fixes#15193
* also remove automatic syscall restart on _kern_select, to match Linux and
BSDs behavior: this fixes parallel build with newer gnu make, which happens
to use pselect.
* also remove automatic syscall restart on _kern_poll.
from https://man7.org/linux/man-pages/man7/signal.7.html
"The following interfaces are never restarted after being
interrupted by a signal handler, regardless of the use of
SA_RESTART; they always fail with the error EINTR when
interrupted by a signal handler: ...
select(2), and pselect(2)."
from https://notes.shichao.io/unp/ch6/
"Berkeley-derived kernels never automatically restart select."
Change-Id: I7f86d221eae1ad93d8a308a75581d2c30a369c9e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3627
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
- Switch to 25MHz clock
- Switch to 4bit transfers mode (the default is 1bit)
Reading and writing SD cards do not seem to work anymore with these
changes. I get invalid data on read, and on write, an interrupt is never
called in some cases.
There are three parts to this change:
- In FetchFileJob, if the request fails with a timeout or IO error
(probably because of unstable connection) attempt to resume the
download with a range request. No limit on number of retries
currently, maybe we should add one.
- In PackageManager, before downloading a file, look around in other
transaction directories in case it's already there. Partial and
complete downloads are differentiated by an attribute which the
fetch file job maintains. For complete downloads, no fetch job is
scheduled, for partial downloads, the fetch job will request the
remainder of the file.
- In BHttpRequest, the implementation of SetRangeStart() and
SetRangeEnd() have been added, along with some refactoring to
handle listener notifications consistently. This also fixed a
bug where the final notification for download progress was not
emitted for compressed data.
Fixes#12414.
Change-Id: I3e285741ed0e5651594a7c2e1c7170644a9d297d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3404
Reviewed-by: Stephan Aßmus <superstippi@gmx.de>
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
On sparc, the minimal page size we can use is 8K. Since B_PAGE_SIZE and
PAGESIZE defines were hardcoded to 4K, this resulted in a lot of
confusion in all code trying to manipulate pages.
- Remove cpu.h from headers/private/kernel/arch/*. It dates back from
NewOS and was not used anymore since our kernel uses B_PAGE_SIZE
(PAGE_SIZE was the only thing defined in this header).
- Add posix/arch/*/limits.h with the arch specific page size and include
it from the main limits.h.
- Adjust bios_ia32/debug.cpp which was the only place using the
PAGE_SIZE constant from the deleted headers.
- Change OS.h to define B_PAGE_SIZE to be the same as POSIX PAGESIZE.
- Define PAGESIZE in the build header if the host OS doesn't.
Change-Id: I8c3732cf952ea3c2f088aa16d216678fbf198b96
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3558
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
The main differences:
- The initialization sequence requires an additional command (this was
already done)
- The layout of the CSD register and the way to compute the device
geometry from it changes
- The read and write commands parameter is a sector number instead of a
byte position
Change-Id: Ie729e333c9748f36b37acd70c970adfd425cf0b6
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3512
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
The SDHCI spec also offers an "advanced DMA" mode where we can use
scatter-gather lists. It would allow to remove several of the DMA
restrictions, but hardware support for it is optional, so we need this
version anyway.
The geometry is retrieved on demand in the first read or write or in a
call to the get geometry or get device size ioctl. It is not possible to
retrieve it from the device initialization because that is called as
part of the mmc_bus scanning, which needs a specific sequence of
commands and keeps the bus locked to prevent drivers to insert their own
commands in the middle of that sequence.
TODO:
- Move the DMA restrictions definition to sdhci_pci and forward it up to
mmc_disk (which is the one creating the IOScheduler)
- Decide if we want to keep non-DMA support (probably should, but it
makes things more complex, because it uses virtual addresses)
Change-Id: Ib1dd14eacf62052d747bfb3ef7820bc5a34d3030
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3471
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
* The "default" of 3MiB wasn't enough for modern larger media
formats, resulting in inability to play 4k video no matter
how much horse power you threw at Haiku. (4k is ~8MiB)
* This dynamically calculates the ChunkCache based on the
video framesize * 2.
* 4k video now plays smoothly on my Ryzen 1800x.
Change-Id: I65bf6bd6fa60ac3196ea70eeeb5e655d43c10bcd
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3598
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
