wireguard-openbsd/gnu/llvm/lib/CodeGen, branch jd/queueboostsWireGuard implementation for the OpenBSD kernelhttps://git.zx2c4.com/wireguard-openbsd/atom/gnu/llvm/lib/CodeGen?h=jd%2Fqueueboosts2020-05-02T05:34:08ZDon't make an illegal adde. Avoids fatal error on PowerPC.2020-05-02T05:34:08Zgkoehlergkoehler@openbsd.org2020-05-02T05:34:08Zurn:sha1:ffca677e9e7ca9efd316fa2f2b6572b193c50cf8
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@
Move the hashed __retguard_* symbols into individual sections and mark2019-11-09T20:29:08Zguentherguenther@openbsd.org2019-11-09T20:29:08Zurn:sha1:0a276a5089a34f31ea48f000c0ffe0023990ca5d
them as COMDATs so that the linker can individually discard them, instead
of just ignoring duplicate symbols but keep the (duplicate) space.
On amd64, this reduces the size of the kernel OPENBSD_RANDOM segment by 82%
and the libc OPENBSD_RANDOM segment by 15%. A port that tb@ is working
on experienced a 97.3% reduction...which let it actually run.
ok mortimer@ deraadt@
Import LLVM 8.0.1 release including clang, lld and lldb.2019-09-01T16:27:59Zpatrickpatrick@openbsd.org2019-09-01T16:27:59Zurn:sha1:b19884b112052f3bf5ff9551374fbd840e2339f2Ensure stack protector slot access is resolved by PEI.2019-08-12T23:43:03Zmortimermortimer@openbsd.org2019-08-12T23:43:03Zurn:sha1:c45c722e0148301536effe267054bb9575edbf35
On arm64, arm, and ppc it is possible that a large stack frame will
cause the stack protector slot to be reallocated at the wrong end of
the frame.
Noticed by tj@. ok patrick@.
Merge LLVM 8.0.0 release.2019-06-23T22:05:08Zpatrickpatrick@openbsd.org2019-06-23T22:05:08Zurn:sha1:f27a781c9811da94fe68b2cf34237c2a61b48c96
Prepared with help from jsg@ and mortimer@
Tested on amd64 by bcallah@, krw@, naddy@
Tested on arm64 by patrick@
Tested on macppc by kettenis@
Tested on octeon by visa@
Tested on sparc64 by claudio@
Import LLVM 8.0.0 release including clang, lld and lldb.2019-06-23T21:36:31Zpatrickpatrick@openbsd.org2019-06-23T21:36:31Zurn:sha1:23f101f37937a1bd4a29726cab2f76e0fb038b35Re-enable RETGUARD leaf function optimization for arm64.2019-05-31T19:15:05Zmortimermortimer@openbsd.org2019-05-31T19:15:05Zurn:sha1:90020ccd6197ab7cf35e5085321274918851be68
It turns out MachineFrameInfo.hasCalls() is unreliable, because it is
up to the backends to update this information whenever they add calls
to a function, and this does not always happen.
ok kettenis@
Do not store the retguard cookie in frame in leaf functions if possible.2019-03-31T02:39:47Zmortimermortimer@openbsd.org2019-03-31T02:39:47Zurn:sha1:52357f02de6027c0408b01b51141d97603ef4460
Makes things slightly faster and also improves security in these functions,
since the retguard cookie can't leak via the stack.
ok deraadt@
Tedu files that got removed in LLVM 7.0.1.2019-01-27T17:11:29Zpatrickpatrick@openbsd.org2019-01-27T17:11:29Zurn:sha1:9f520fd6abbfde117cf551fb4eacdd2687a606a9Merge LLVM 7.0.1 release.2019-01-27T16:55:35Zpatrickpatrick@openbsd.org2019-01-27T16:55:35Zurn:sha1:5072e66b18f5d3ac6abf1c18a092e85eab2d7568
With fixes from mortimer@ (thanks!)
Tested by many, especially naddy@ (thanks!)