Summary:
To be noted, this pattern is not unhandled by instcombine per-se,
it is somehow does end up being folded when one runs opt -O3,
but not if it's just -instcombine. Regardless, that fold is
indirect, depends on some other folds, and is thus blind
when there are extra uses.

This does address the regression being exposed in D63992.

https://godbolt.org/z/7DGltU
https://rise4fun.com/Alive/EPO0

Fixes [[ https://bugs.llvm.org/show_bug.cgi?id=42459 | PR42459 ]]

Reviewers: spatel, nikic, huihuiz

Reviewed By: spatel

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63993

llvm-svn: 364792

Summary:
Given pattern:
`icmp eq/ne (and ((x shift Q), (y oppositeshift K))), 0`
we should move shifts to the same hand of 'and', i.e. rewrite as
`icmp eq/ne (and (x shift (Q+K)), y), 0`  iff `(Q+K) u< bitwidth(x)`

It might be tempting to not restrict this to situations where we know
we'd fold two shifts together, but i'm not sure what rules should there be
to avoid endless combine loops.

We pick the same shift that was originally used to shift the variable we picked to shift:
https://rise4fun.com/Alive/6x1v

Should fix [[ https://bugs.llvm.org/show_bug.cgi?id=42399 | PR42399]].

Reviewers: spatel, nikic, RKSimon

Reviewed By: spatel

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63829

llvm-svn: 364791

llvm-svn: 364790

llvm-svn: 364789

Summary:
The stride should depend on the wave size, not the hardware generation.

Also, the 32_FLOAT format is 0x16, not 16; though that shouldn't be
relevant.

Change-Id: I088f93bf6708974d085d1c50967f119061da6dc6

Reviewers: arsenm, rampitec, mareko

Subscribers: kzhuravl, jvesely, wdng, yaxunl, dstuttard, tpr, t-tye, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63808

llvm-svn: 364788

This is easy to handle and avoids legalization artifacts which are
likely to obscure combines.

llvm-svn: 364787

llvm-svn: 364786

Summary:
Now we store the errors for the Decls in the "to" context too. For
that, however, we have to put these errors in a shared state (among all
the ASTImporter objects which handle the same "to" context but different
"from" contexts).

After a series of imports from different "from" TUs we have a "to" context
which may have erroneous nodes in it. (Remember, the AST is immutable so
there is no way to delete a node once we had created it and we realized
the error later.) All these erroneous nodes are marked in
ASTImporterSharedState::ImportErrors.  Clients of the ASTImporter may
use this as an input. E.g. the static analyzer engine may not try to
analyze a function if that is marked as erroneous (it can be queried via
ASTImporterSharedState::getImportDeclErrorIfAny()).

Reviewers: a_sidorin, a.sidorin, shafik

Subscribers: rnkovacs, dkrupp, Szelethus, gamesh411, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D62376

llvm-svn: 364785

isVCC has the same bug, but isn't used in a context where it can cause
a problem.

llvm-svn: 364784

llvm-svn: 364783

llvm-svn: 364782

Set global enable bits (i.e. bits 1, 3, 5, 7) to enable watchpoints
on NetBSD rather than the local enable bits (0, 2, 4, 6).  The former
are necessary for watchpoints to be correctly recognized by the NetBSD
kernel.  The latter cause them to be reported as trace points.

Differential Revision: https://reviews.llvm.org/D63792

llvm-svn: 364781

Fix the watchpoint/breakpoint code to search for matching thread entry
in m_threads explicitly rather than assuming that it will be present
at specified index.  The previous code segfault since it wrongly assumed
that the index will match LWP ID which was incorrect even for a single
thread (where index was 0 and LWP ID was 1).

While fixing that off-by-one error would help for this specific task,
I believe it is better to be explicit in what we are searching for.

Differential Revision: https://reviews.llvm.org/D63791

llvm-svn: 364780

Provide a (conditional) support for the new PT_GETXSTATE
and PT_SETXSTATE ptrace() requests, and use them to implement getting
and setting YMM registers.  The functions used for splitting
and recombining YMM register data are based on matching functions
in FreeBSD plugin, with some simplification and updates to match NetBSD
structures.

Differential Revision: https://reviews.llvm.org/D63545

llvm-svn: 364779

Fix stack-use-after-scope errors from r364512. One instance was already
fixed in r364611 - this patch simplifies that fix and addresses one more
instance of similar code.

Discussed in: https://reviews.llvm.org/D63905

llvm-svn: 364778

Summary:
LLVM issues a warning if passed unknown target features. Neither I nor
@asb noticed this until after https://reviews.llvm.org/D63498 landed.

This patch stops passing the (unknown) "save-restore" target feature to
the LLVM backend, but continues to emit a warning if a driver asks for
`-msave-restore`. The default of assuming `-mno-save-restore` (and
emitting no warnings) remains.

Reviewers: asb

Reviewed By: asb

Subscribers: rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, kito-cheng, shiva0217, jrtc27, zzheng, edward-jones, rogfer01, MartinMosbeck, brucehoult, the_o, rkruppe, PkmX, jocewei, psnobl, benna, Jim, cfe-commits, asb

Tags: #clang

Differential Revision: https://reviews.llvm.org/D64008

llvm-svn: 364777

Now that r364751 has been reverted, we need to revert this fixup
as well.

llvm-svn: 364776

Summary:
SCRUB_LOOP_COMMENT_RE was introduced in https://reviews.llvm.org/D31285
This works for some loops.

However, we may generate lines with loop comments only.
And since we don't scrub leading white spaces, this will leave an empty
line there, and FileCheck will complain it.

eg: llvm/test/CodeGen/PowerPC/PR35812-neg-cmpxchg.ll:27:15:
error: found empty check string with prefix 'CHECK:'
; CHECK-NEXT:

This prevented us from using the `update_llc_test_checks.py` for quite some cases.

We should still keep the comment token there, so that we can safely
scrub the loop comment without breaking FileCheck.

Reviewers: timshen, hfinkel, lebedev.ri, RKSimon

Subscribers: nemanjai, jfb, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63957

llvm-svn: 364775

llvm-svn: 364774

This fixes a failing testcase on Fedora 30 x86_64 (regression Fedora 29->30):

PASS:
./bin/lldb ./lldb-test-build.noindex/functionalities/unwind/noreturn/TestNoreturnUnwind.test_dwarf/a.out -o 'settings set symbols.enable-external-lookup false' -o r -o bt -o quit
  * frame #0: 0x00007ffff7aa6e75 libc.so.6`__GI_raise + 325
    frame #1: 0x00007ffff7a91895 libc.so.6`__GI_abort + 295
    frame #2: 0x0000000000401140 a.out`func_c at main.c:12:2
    frame #3: 0x000000000040113a a.out`func_b at main.c:18:2
    frame #4: 0x0000000000401134 a.out`func_a at main.c:26:2
    frame #5: 0x000000000040112e a.out`main(argc=<unavailable>, argv=<unavailable>) at main.c:32:2
    frame #6: 0x00007ffff7a92f33 libc.so.6`__libc_start_main + 243
    frame #7: 0x000000000040106e a.out`_start + 46

vs.

FAIL - unrecognized abort() function:
./bin/lldb ./lldb-test-build.noindex/functionalities/unwind/noreturn/TestNoreturnUnwind.test_dwarf/a.out -o 'settings set symbols.enable-external-lookup false' -o r -o bt -o quit
  * frame #0: 0x00007ffff7aa6e75 libc.so.6`.annobin_raise.c + 325
    frame #1: 0x00007ffff7a91895 libc.so.6`.annobin_loadmsgcat.c_end.unlikely + 295
    frame #2: 0x0000000000401140 a.out`func_c at main.c:12:2
    frame #3: 0x000000000040113a a.out`func_b at main.c:18:2
    frame #4: 0x0000000000401134 a.out`func_a at main.c:26:2
    frame #5: 0x000000000040112e a.out`main(argc=<unavailable>, argv=<unavailable>) at main.c:32:2
    frame #6: 0x00007ffff7a92f33 libc.so.6`.annobin_libc_start.c + 243
    frame #7: 0x000000000040106e a.out`.annobin_init.c.hot + 46

The extra ELF symbols are there due to Annobin (I did not investigate why this problem happened specifically since F-30 and not since F-28).
It is due to:

Symbol table '.dynsym' contains 2361 entries:
Valu e          Size Type   Bind   Vis     Name
0000000000022769   5 FUNC   LOCAL  DEFAULT _nl_load_domain.cold
000000000002276e   0 NOTYPE LOCAL  HIDDEN  .annobin_abort.c.unlikely
...
000000000002276e   0 NOTYPE LOCAL  HIDDEN  .annobin_loadmsgcat.c_end.unlikely
...
000000000002276e   0 NOTYPE LOCAL  HIDDEN  .annobin_textdomain.c_end.unlikely
000000000002276e 548 FUNC   GLOBAL DEFAULT abort
000000000002276e 548 FUNC   GLOBAL DEFAULT abort@@GLIBC_2.2.5
000000000002276e 548 FUNC   LOCAL  DEFAULT __GI_abort
0000000000022992   0 NOTYPE LOCAL  HIDDEN  .annobin_abort.c_end.unlikely

Differential Revision: https://reviews.llvm.org/D63540

llvm-svn: 364773

As discussed in https://reviews.llvm.org/D63829
*if* *both* shifts are one-use, we'd most likely want to produce `lshr`,
and not rely on ordering.

Also, there should likely be a *separate* fold to do this reordering.

llvm-svn: 364772

Summary:
During import of a specific Decl D, it may happen that some AST nodes
had already been created before we recognize an error. In this case we
signal back the error to the caller, but the "to" context remains
polluted with those nodes which had been created. Ideally, those nodes
should not had been created, but that time we did not know about the
error, the error happened later.  Since the AST is immutable (most of
the cases we can't remove existing nodes) we choose to mark these nodes
as erroneous.
Here are the steps of the algorithm:
1) We keep track of the nodes which we visit during the import of D: See
ImportPathTy.
2) If a Decl is already imported and it is already on the import path
(we have a cycle) then we copy/store the relevant part of the import
path. We store these cycles for each Decl.
3) When we recognize an error during the import of D then we set up this
error to all Decls in the stored cycles for D and we clear the stored
cycles.

Reviewers: a_sidorin, a.sidorin, shafik

Subscribers: rnkovacs, dkrupp, Szelethus, gamesh411, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D62375

llvm-svn: 364771

Add code to catch pattern for commutative instructions for VLCR.

Patch by Suyog Sarda.

llvm-svn: 364770

llvm-svn: 364769

llvm-svn: 364768

llvm-svn: 364767

llvm-svn: 364766

Select s64 eq/ne scalar icmp.

llvm-svn: 364765

So we indeed to have this fold, but only if +1 is not the last operation..

llvm-svn: 364764

llvm-svn: 364763

llvm-svn: 364762

This was checking the size of the register with the value of the size,
which happens to be exec. Also fix assuming VCC is 64-bit to fix
wave32.

Also remove some untested handling for physical registers which is
skipped. This doesn't insert the V_CNDMASK_B32 if SCC is the physical
copy source. I'm not sure if this should be trying to handle this
special case instead of dealing with this in copyPhysReg.

llvm-svn: 364761

Zext from s1 is the only case where this should do anything with the
current legal extensions.

llvm-svn: 364760

llvm-svn: 364759

llvm-svn: 364758

llvm-svn: 364757

llvm-svn: 364756

llvm-svn: 364755

A bunch of places were checking that DataBufferHeap::GetBytes returns a
non-null pointer right after constructing it. The only time when
GetBytes returns a null pointer is if it is empty (and I'm not sure that
even this is a good idea), but that is clearly not the case here, as the
buffer was constructed with a non-zero size just a couple of lines back.

llvm-svn: 364754

llvm-svn: 364753