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 will probably be just stubs for the significant future, but,
here it is anyway.
Regarding the naming: Yes, the official name is "aarch64." However,
Linux, FreeBSD, and Zircon all call it "arm64", and so we will do the same.
I've configured it initially to be a Clang-only port, making no
changes to GCC buildtools whatsoever here. We'll see if that sticks,
however.
Check a range of characters to see if they have the same width.
Change-Id: I9cef12f78d1e1d01acc6284eae90dbfb86166fc6
Reviewed-on: https://review.haiku-os.org/47
Reviewed-by: Adrien Destugues <pulkomandy@gmail.com>
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
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.
Since they were being Use'd at global scope, the FreeBSD versions
of a number of POSIX headers were being used instead of the POSIX ones,
which breaks the build on non-x86 as these headers presently do not work
elsewhere.
Instead, just include the base directory, and then include the compat
headers manually in the one place that actually requires them.
Also fix all the other files that expected errno.h and others to
be included implicitly, which they now are not.
This should fix the PowerPC and (part of the) ARM builds.
Previously we initialized variables and ran a few $CC tests first
(which was what the old version of --update needed.) Now, we can do it
almost immediately after the script begins.
Spotted by running --update on a GCCless system (as the $CC tests
gave warning messages due to no GCC, while CC= was set in --update
environs.)
Clang's equivalent of .tfloat is .x86_fp80, but apparently that
is not enabled for Haiku. So for now, use .double instead.
This is the same thing MinGW does.
This block was originally disabled in libbind, and we enabled it
independently; and so did NetBSD. But they also made one other
fix:
From: christos <christos@netbsd.org>
Date: Thu, 28 Dec 2017 15:12:15 +0000
Subject: [PATCH] PR/52837: Michael Kaufmann: getaddrinfo() resolves "127.0.0.1
www.example.com" to 127.0.0.1. Apply the patch from FreeBSD and explain the
rationale.
I attempted to merge NetBSD's new irs code, but their changes turned
out to be too invasive to do trivially, but in the process I had to
read our commit logs to find what we've changed, so at least I can
mark that.
Commits merged from the semi-official Git mirror of NetBSD
trunk (https://github.com/IIJ-NetBSD/netbsd-src/).
Commit authors/messages in chronological order follow:
---------------------------------------
From: christos <christos@netbsd.org>
Date: Thu, 28 Sep 2017 23:32:01 +0000
Subject: [PATCH] PR/52578: Benjamin M. Schwartz Set the AD bit when DNSSEC is
enabled (RFC 6840 Section 5.7).
From: lukem <lukem@netbsd.org>
Date: Fri, 2 Mar 2018 06:31:53 +0000
Subject: [PATCH] PR/48585: Set errno when returning NULL for AF_INET
In inet_ntop4(), errno wasn't set before returning NULL.
Seems like an oversight in the existing fix for PR/48585.
Noticed by code inspection.
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
* <x86_gcc2>runtime_loader can also handle x86 binaries.
* At the moment the x86_64 kernel will execute the x86 runtime_loader
when it sees x86 and x86_gcc2 binaries. The reason is checking the GCC
version of a binary in the kernel seems cumbersome, the kernel should
just fallback using the x86_gcc2 runtime_loader when the x86 one is missing.
We lost these tunings when I moved us away from board focused
builds. I feel like most of our ARM interest is around ARMv7+
Change-Id: Ie301d275a74d48ee3d0c4c7dc7d6cdd635288a7b
* gcc 7.x defines __arm__ and __ARM__ (and others)
* clang defines __arm__ and __arm
* cleanup a few related ifdef vs if macros
Change-Id: I5da4bafac590f6fa3e10e543688001c2449f840d
* B_MEDIA_SEEKABLE was once a flag with one bit set. Due
to the complexity of streaming mediums, I had to split
this flag into B_MEDIA_SEEK_BACKWARD and B_MEDIA_SEEK_FORWARD
and never noticed that this broke the flag check.
* Reported by Murai Takashi through gcc8.
This requires a trunk build of Clang (the flag was only implemented
& introduced 12 days ago), but at present, full builds will fail
due to an unrelated Clang bug: https://bugs.llvm.org/show_bug.cgi?id=38356
* __NO_INLINE__ fixes the cross-build on some glibc-based systems with
newer compilers, as it prevents glibc from declaring functions inline
that we override in libroot_build.
* We can now enable tree-vrp as long as no-delete-null-pointer-checks
goes where it used to.