The kernel heap only uses object caches for objects up to size 8192.
Larger allocations have to go through the raw allocator. That can
get pretty expensive.
Adding instrumentation around the malloc/free calls in this function
showed that on my machine, some 596ms during boot were spent on
*malloc/free alone*, all else aside. After this change, we are at
around 110ms, or a >5x improvement. Running an fgrep -R on /system/
after boot increased the cumulative time in memory functions to over
5 seconds, while after this change it is "only" 1170ms.
Honestly, it seems like the object depots should be able to be faster
than that, even if this function is called thousands of times. But that
is a problem for a different investigation.
It would be even faster for every consumer of this data in
packagefs just allocated one set of buffers up front, or at least
for a single "read session", but plumbing that all the way
through the myriad abstractions of the Package Kit will
not be easy, and is left for another time, as well.
While the BSDs and glibc seem to have various _r functions,
they all return int for errors instead of a pointer, making ours
exactly backwards of theirs for error reporting and thus useless.
So, remove them from the header entirely. They are left in
for ABI backwards compatibility for the time being.
A few constants not used by anything in the tree (i.e. not actually
implemented by libnetwork/netresolv) were also dropped. Some deprecated
or non-standard functions were placed behind _DEFAULT_SOURCE or deleted
entirely.
The header itself is now organized approximately as the BSDs do,
although with various Haiku-isms instead of BSD-isms where appropriate.
This cuts out a lot of unnecessary protocol entires found in the
longer NetBSD (and originally from ISC) lists. This data is used
in the implementation of getservent(), etc.
We shouldn't be doing these crazy things.
The only usage of this is the NextID rule which computes an unique
number by incrementing a variable. This is used for example to generate
names for the extra files used in xattr emulation.
But we certainly don't need multiplication, division, etc. That's better
done in scripts (Python, for example) in rule actions. So keep just
enough of this for NextID to work, and we could probably make it even
simpler.
Change-Id: Id06d4f5e236205203fe8d604de440753b63801bc
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6088
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Tested-by: Automation <automation@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Saves at least a thread and a few semaphores per app.
Change-Id: Ied43e5944471918b34b897722346a25d159090df
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6073
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Move common add/remove code to the base class and improve encapsulation.
Loading FreeType is the responsibility of the global font manager.
Rename FontManagerBase back to FontManager.
Change-Id: I9396112f01583e19d03011106fd1fa1e37958ee8
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6072
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
App-loaded fonts get the same IDs for different apps, so the cache
entries can be easily overwritten or used for an unexpected font. The
FontManager adds a unique chunk to the string.
Change-Id: I67ea9971784f425370d1aad516af4dae8489c45b
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6071
Tested-by: Automation <automation@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
* Record the read-only status of a partition in settings/mount_server
so it will be applied if automatically mounting on next boot.
* Fixes#18081.
Change-Id: Idd7267fd356f35689bb6ad79de4a45bf071164f1
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6080
Tested-by: Automation <automation@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
* We depend on FDT passed to the bootloader now
from EFI or u-boot via fdt_addr_r now
* We leave an fdt path within the boot partition
(for now) to allow / encourage users to optionally
plug in their own DTB's for troubleshooting. (only
on u-boot loader)
Change-Id: I3f2d81b60d46f388f333d5caa27aa77e6e36447d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6081
Reviewed-by: Alex von Gluck IV <kallisti5@unixzen.com>
Reviewed-by: X512 <danger_mail@list.ru>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Tested-by: Automation <automation@haiku-os.org>
In fixed size mode, the menu field always uses all the available width.
In non-fixed size mode, the menu field will resize itself to be as small
as possible.
With layout, usually the difference isn't noticeable, since the layout
will already try to resize the control to the smallest possible size.
But there are a few cases where it makes a difference, when the layout
is over-constrained and the menu field can't be made as small as
possible. In that case, the menu field would be forced to fill its
allocated space, where we can instead make it a little smaller.
Change-Id: I911d497218a09aab3824865968558df5d4b3cf98
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6076
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Reviewed-by: John Scipione <jscipione@gmail.com>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
These were introduced by mistake during the addition of the layout
system. They have been private since 2011. It seems safe to assume that
nothing uses them anymore and they can now be removed safely.
Change-Id: I98d030096f9cb06fccc25233fe4da17d0213050e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6075
Tested-by: Automation <automation@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
The code is endian-dependant because the alpha channel is not at the
same place. This resulted in transparent white in icons being converted
to partially transparent yellow when attempting to darken the icons, for
example.
Change-Id: I57916139ba37b8d7ae0f9e9ba9e74c1d496d7b05
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6077
Tested-by: Automation <automation@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
The previous logic of _AddTrim plus _TrimNext meant that if a trim
range was specified in the final call to _TrimNext (i.e. with force=true),
but the trim buffer was already full at that point, that last range
would not actually be trimmed.
Now, _AddTrim returns true when the buffer is filled and trimming
should be done, rather than delaying it and requiring a second
add after trimming.
Change-Id: I4b782948e8dc9267c63e61bce0c078fd4e834177
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6070
Tested-by: Automation <automation@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
BSD extension to set both the input and output speed of a termios
structure.
Fixes#18220.
Change-Id: I8c4a06b4be4aa55b8ce35cb7f62552fc47a8e8d0
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6049
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
There's only one implementation of smix so we don't
need any selection logic.
Change-Id: Id43e4ed49ee5a96755c3c4ab975bd128024e57b0
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6065
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
Define a dedicated OnDiskData structure for each on-disk structure. This
must match the on-disk layout, except for endianness, which is handled
by _SwapEndian methods. These structure are "plain old data" so we can
use offsetof on them. They are wrapped in an easier to use C++ API.
This resolves a lot of problems with the previous code: warnings caused
by the use of offsetof as well as a much simpler instanciation of the
objects from on-disk data.
Also fixed another problem with UUIDs, where the UUIDs were handled by
pointers in a lot of place where it was not necessary. Use references
instead. The V4 structures which don't have an UUID will return a "null"
(zero-filled) one.
Change-Id: Ifb2bf6ab94906ca50410dd3446d3566615392ca2
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6021
Reviewed-by: Raghav Sharma <raghavself28@gmail.com>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
..instead of readv_pos/writev_pos. This way, we can be sure that we
are remapping them properly even under Haiku, as well as remove some
potential confusions.
bfs_shell seems to still work fine on a regular build.
All drivers using IORequest should have any problems here already caught
by it, and overruns in other drivers would overrun the buffer and cause
corruptions, so we should no longer silently be ignoring this,
presuming it is still a problem at all.
As the TODO comment indicated that this logic could also be used for
writes if this wasn't a problem, I have added a separate TODO for that.
They are not about reading and writing pages, but just iovecs.
As the passed iovecs use void*s, it can't possibly be working with
physical pages on 32-bit systems with PAE.
It appears nothing uses or implements these functions anyway,
as there was nothing else in the tree I had to adjust after making
this change...
This is required to make crypt_r() really reentrant.
Change-Id: I81497be16cae21fabb5895a88e27e6c20d3e06e0
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6066
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
On adding a font, we may or not have other faces of the same family, so
we may be trying to add the new face to an already managed family. In
that case, we can't delete it when we get an error inserting the new
style.
This same code was already in place before the user fonts, but it's much
more unlikely to trigger, except maybe on purpose. A later patch
refactoring a bunch of things will join both into the base class.
Fixes#18169
Change-Id: I78053b6e941c18b952fc2bd64f393e26d3d53b0a
* Enable user to drag the WebPositive web page icon to the bookmark
bar to create a bookmark there, or drag it to a Tracker window to
create a bookmark in the displayed directory.
* Send a message to the drag target that can be handled by Tracker's
generic drop routine.
* Overload _CreateBookmark with a more flexible version and a
_CreateBookmark(BMessage*) that handles both the icon being dropped
on the bookmark bar, and the message that Tracker sends if the icon
is dropped on Tracker.
* Account for the fact that, when _CreateBookmark(BMessage*) is called,
Tracker may or may not have already determined the file name to use and
created the file, depending on whether the icon was dragged to
Tracker or not.
* Use page-specific small and large icons for the bookmark file, if
they are available (currently Haiku WebKit doesn't seem to provide
them though).
* Follows CharacterMap as a model for dragging an icon and ShowImage
for dragging to Tracker.
* Fixes#10795.
Change-Id: I7f32013cc1372dab1894b5d92335d3a4cbfb671f
Reviewed-on: https://review.haiku-os.org/c/haiku/+/6007
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
