Age | Commit message (Collapse) | Author |
|
|
|
ok deraadt@
|
|
Lost in the cleanup.
Noticed by deraadt@
|
|
powerpc. This matched what FreeBSD does.
ok patrick@
|
|
This was originally done so that other things in-tree such as vi could link
to libperl and to support non-shared vax.
This brings us more in-line with upstream and makes things more understandable.
It now links libperl.so to libm, which some software needs, and stops building
and installing libperl.a.
OK bluhm@
|
|
|
|
32-bit PowerPC doesn't have instructions for lock-free atomic ops on
8-byte values, and needs libcalls like __atomic_fetch_add_8(). In
code like "_Atomic long long a; a++;", clang doesn't emit a libcall.
This was causing linker errors on symbols like __sync_fetch_and_add_8.
Now that LLVM knows the max atomic size, its AtomicExpandPass changes
these 8-byte ops into libcalls.
ok mortimer@
|
|
All functional changes were committed already, this updates the
version number, Module::Corelist, and documentation.
|
|
|
|
Found by: ManhND of The Tarantula Team, VinCSS (a member of Vingroup),
Hugo van der Sanden, Slaven Rezic, and Sergey Aleynikov
Fixed by: John Lightsey, Hugo van der Sanden, and Karl Williamson
Addresses:
* CVE-2020-10543
* CVE-2020-10878
* CVE-2020-12723
|
|
generate fatal compiler warnings at least when cross-compiling and
fixing the code isn't trivial.
ok patrick@, drahn@
|
|
This should simplify bringup and make it easier to support Big Endian
and Little Endian with the same code.
May be reconsidered if it causes too many problems with Ports.
ok kettenis@
|
|
ok kettenis@
|
|
Disable PIC/PIE for powerpc64 for now.
|
|
"where is the kaboom?" deraadt@
|
|
ok hackroom@
|
|
arithmetics on NULL, like we did in r1.21 on elflink.c, to stop
clang from complaining when setting up a cross-toolchain.
ok kettenis@
|
|
clang defaulted to -fomit-frame-pointer when optimizing PowerPC code
for Linux and NetBSD. Now do it for all systems, including OpenBSD.
This affects both 32-bit and powerpc64 code.
ok kettenis@
|
|
This lets the kernel detect retguard traps and send SIGABRT instead
of SIGEMT.
SIGEMT does not indicate correctly the nature of the error (stack
overflow, violation of control flow). It can confuse the user to restart
the program without further investigation.
Prompted by and OK deraadt@
OK mortimer@
|
|
Prompted and tested by drahn@
ok deraadt@ drahn@ kettenis@ visa@
|
|
When the DAG truncates an ISD::ADDE node, DAGCombiner may optimize it
by making an adde with smaller operands. PowerPC has i1 registers,
and may truncate an i32 adde to i1, but an i1 adde is not legal for
PowerPC, and the legalize-ops phase can't fix it. This was causing
"fatal error: error in backend: Cannot select..."
cwen@ reported the error
ok mortimer@ kettenis@ deraadt@
|
|
ok hackroom@
|
|
Avoids using the installed Cwd module while building perl
|
|
"hard-quad-float" feature is available. Add missing replacement
instruction patterns that are needed to emit alternative code for
conditional moves of quad-precision floats.
ok mortimer@
|
|
ok hackroom@
|
|
symbols that were hidden. Fixes building Mesa on hppa.
ok deraadt@
|
|
Missed somehow in perl update
|
|
https://metacpan.org/pod/release/SHAY/perl-5.30.2/pod/perldelta.pod
Incompatible Changes
There are no changes intentionally incompatible with 5.30.0.
Updated Modules and Pragmata
* Compress::Raw::Bzip2 has been upgraded from version 2.084 to 2.089.
* Module::CoreList has been upgraded from version 5.20191110 to 5.20200314.
Selected Bug Fixes
* printf() or sprintf() with the %n format no longer cause a panic
on debugging builds, or report an incorrectly cached length value
when producing SVfUTF8 flagged strings.
* A memory leak in regular expression patterns has been fixed.
* A read beyond buffer in grok_infnan has been fixed.
* An assertion failure in the regular expression engine has been fixed.
* (?{...}) eval groups in regular expressions no longer unintentionally
trigger "EVAL without pos change exceeded limit in regex".
Proceed when you feel comfortable. deraadt@
|
|
|
|
|
|
other clang platforms do.
ok jca@
|
|
"where is the kaboom?" deraadt@
|
|
ok mortimer@
|
|
ok hackroom@
|
|
ok bluhm@
|
|
OpenBSD macppc builds these libs with clang. Their ABI is changing
because of clang -msvr4-struct-return
|
|
Add these options from gcc to clang:
-maix-struct-return # return all structs in memory
-msvr4-struct-return # return small structs in r3/r4
ok mortimer@ jca@ deraadt@
|
|
ok hackroom@
|
|
which means we do stop creating other jobs as soon as we run something looking
like "make" to avoid unwanted recursion, so you would never hit that.
But in binutils-2.17, the developers changed all/info to do recursive makes
with some of the same dependencies. Calling both at the same time becomes
an obvious race, and should be fixed for no other reason than correctness.
okay guenther@
|
|
Timing is good deraadt@, OK sthen@
|
|
Timing is good deraadt@, OK sthen@
|
|
Timing is good deraadt@, OK sthen@
|
|
Timing is good deraadt@, OK sthen@
|
|
OK espie@ sthen@ deraadt@
|
|
|
|
|
|
ok bluhm kettenis
|
|
okay millert@, tb@
|
|
"Go for it" kettenis@
|
|
this got fixed in recent binutils, but it's GPLv3.
So do the same thing in a slightly different way
okay guenther@
|