> Date: Tue, 9 Apr 2024 15:18:10 -0500 > From: Segher Boessenkool <seg...@kernel.crashing.org>
> All (target-specific) new testsuite failures are just like that: bad > testcases! With a touch of bad assumptions by port-specific code, no doubt. Maybe also rtx costs including my pet peeve, the default implementation of insn_costs (the one that doesn't look at the destination of setters and which when you try fixing it, pulls you down a rabbit-hole of cost-related regressions that even Bernd S. backed away from). > So no, no reversion. (...) > > That's the only test that's improved to the point of > > affecting test-patterns. E.g. pr93372-5.c (which references > > pr93372-2.c) is also improved, though it retains a redundant > > compare insn. (PR 93372 was about regressions from the cc0 > > representation; not further improvement like here, thus it's > > not tagged. Though, I did not double-check whether this > > actually *was* a regression from cc0.) > > Interesting that this improved tests for you. Huh. Do you have an > explanation how this happened? Just a hunch: less combine churn (more straightforward code) made cmpelim's job easier, same thing you wrote in order words: > I suspect that as uaual it is just a > side effect of random factors: combine is opportunistic, always does the > first change it thinks good, not considering what this then does for > other possible combinations; it is greedy. It would be nice to see > written out what happens in this example though :-) Yes it would, but I have other things on my plate. Besides, it's your patch, can't rob you of the fun. I committed the revert below, but hope to re-apply (re-revert) it in stage 1, when as per Richard B's message the combine improvement will reappear. brgds, H-P -- >8 -- From: Hans-Peter Nilsson <h...@axis.com> Date: Wed, 10 Apr 2024 17:24:10 +0200 Subject: [PATCH] Revert "testsuite/gcc.target/cris/pr93372-2.c: Handle xpass from combine improvement" This reverts commit 4c8b3600c4856f7915281ae3ff4d97271c83a540. --- gcc/testsuite/gcc.target/cris/pr93372-2.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) diff --git a/gcc/testsuite/gcc.target/cris/pr93372-2.c b/gcc/testsuite/gcc.target/cris/pr93372-2.c index 2ef6471a990b..912069c018d5 100644 --- a/gcc/testsuite/gcc.target/cris/pr93372-2.c +++ b/gcc/testsuite/gcc.target/cris/pr93372-2.c @@ -1,20 +1,19 @@ /* Check that eliminable compare-instructions are eliminated. */ /* { dg-do compile } */ /* { dg-options "-O2" } */ -/* { dg-final { scan-assembler-not "\tcmp|\ttest" } } */ -/* { dg-final { scan-assembler-not "\tnot" } } */ -/* { dg-final { scan-assembler-not "\tlsr" } } */ -/* We should get just one move, storing the result into *d. */ -/* { dg-final { scan-assembler-times "\tmove" 1 } } */ +/* { dg-final { scan-assembler-not "\tcmp|\ttest" { xfail *-*-* } } } */ +/* { dg-final { scan-assembler-not "\tnot" { xfail cc0 } } } */ +/* { dg-final { scan-assembler-not "\tlsr" { xfail cc0 } } } */ int f(int a, int b, int *d) { int c = a - b; - /* We used to get a cmp.d with the original operands here. */ + /* Whoops! We get a cmp.d with the original operands here. */ *d = (c == 0); - /* We used to get a suboptimal sequence, but now we get the optimal "sge" - (a.k.a "spl") re-using flags from the subtraction. */ + /* Whoops! While we don't get a test.d for the result here for cc0, + we get a sequence of insns: a move, a "not" and a shift of the + subtraction-result, where a simple "spl" would have done. */ return c >= 0; } -- 2.30.2 brgds, H-P