diff options
author | Richard Biener <rguenther@suse.de> | 2024-01-15 12:55:20 +0100 |
---|---|---|
committer | Richard Biener <rguenther@suse.de> | 2024-01-23 08:08:27 +0100 |
commit | a98d5130a6dcff2ed4db371e500550134777b8cf (patch) | |
tree | acb49438fccc48092005a62035b75ed384574191 | |
parent | 7218f5050cb7163edae331f54ca163248ab48bfa (diff) | |
download | gcc-a98d5130a6dcff2ed4db371e500550134777b8cf.zip gcc-a98d5130a6dcff2ed4db371e500550134777b8cf.tar.gz gcc-a98d5130a6dcff2ed4db371e500550134777b8cf.tar.bz2 |
rtl-optimization/113255 - base_alias_check vs. pointer difference
When the x86 backend generates code for cpymem with the rep_8byte
strathegy for the 8 byte aligned main rep movq it needs to compute
an adjusted pointer to the source after doing a prologue aligning
the destination. It computes that via
src_ptr + (dest_ptr - orig_dest_ptr)
which is perfectly fine. On RTL this is then
8: r134:DI=const(`g'+0x44)
9: {r133:DI=frame:DI-0x4c;clobber flags:CC;}
REG_UNUSED flags:CC
56: r129:DI=const(`g'+0x4c)
57: {r129:DI=r129:DI&0xfffffffffffffff8;clobber flags:CC;}
REG_UNUSED flags:CC
REG_EQUAL const(`g'+0x4c)&0xfffffffffffffff8
58: {r118:DI=r134:DI-r129:DI;clobber flags:CC;}
REG_DEAD r134:DI
REG_UNUSED flags:CC
REG_EQUAL const(`g'+0x44)-r129:DI
59: {r119:DI=r133:DI-r118:DI;clobber flags:CC;}
REG_DEAD r133:DI
REG_UNUSED flags:CC
but as written find_base_term happily picks the first candidate
it finds for the MINUS which means it picks const(`g') rather
than the correct frame:DI. This way find_base_term (but also
the unfixed find_base_value used by init_alias_analysis to
initialize REG_BASE_VALUE) performs pointer analysis isn't
sound. The following restricts the handling of multi-operand
operations to the case we know only one can be a pointer.
This for example causes gcc.dg/tree-ssa/pr94969.c to miss some
RTL PRE (I've opened PR113395 for this). A more drastic patch,
removing base_alias_check results in only gcc.dg/guality/pr41447-1.c
regressing (so testsuite coverage is bad). I've looked at
gcc.dg/tree-ssa tests and mostly scheduling changes are present,
the cc1plus .text size is only 230 bytes worse. With the this
less drastic patch below most scheduling changes are gone.
x86_64 might not the very best target to test for impact, but
test coverage on other targets is unlikely to be very much better.
PR rtl-optimization/113255
* alias.cc (find_base_term): Remove PLUS/MINUS handling
when both operands are not CONST_INT_P.
* gcc.dg/torture/pr113255.c: New testcase.
-rw-r--r-- | gcc/alias.cc | 28 | ||||
-rw-r--r-- | gcc/testsuite/gcc.dg/torture/pr113255.c | 27 |
2 files changed, 32 insertions, 23 deletions
diff --git a/gcc/alias.cc b/gcc/alias.cc index b2ec480..6fad4b2 100644 --- a/gcc/alias.cc +++ b/gcc/alias.cc @@ -2077,31 +2077,13 @@ find_base_term (rtx x, vec<std::pair<cselib_val *, if (tmp1 == pic_offset_table_rtx && CONSTANT_P (tmp2)) return find_base_term (tmp2, visited_vals); - /* If either operand is known to be a pointer, then prefer it - to determine the base term. */ - if (REG_P (tmp1) && REG_POINTER (tmp1)) - ; - else if (REG_P (tmp2) && REG_POINTER (tmp2)) - std::swap (tmp1, tmp2); - /* If second argument is constant which has base term, prefer it - over variable tmp1. See PR64025. */ - else if (CONSTANT_P (tmp2) && !CONST_INT_P (tmp2)) + if (CONST_INT_P (tmp1)) std::swap (tmp1, tmp2); - /* Go ahead and find the base term for both operands. If either base - term is from a pointer or is a named object or a special address - (like an argument or stack reference), then use it for the - base term. */ - rtx base = find_base_term (tmp1, visited_vals); - if (base != NULL_RTX - && ((REG_P (tmp1) && REG_POINTER (tmp1)) - || known_base_value_p (base))) - return base; - base = find_base_term (tmp2, visited_vals); - if (base != NULL_RTX - && ((REG_P (tmp2) && REG_POINTER (tmp2)) - || known_base_value_p (base))) - return base; + /* We can only handle binary operators when one of the operands + never leads to a base value. */ + if (CONST_INT_P (tmp2)) + return find_base_term (tmp1, visited_vals); /* We could not determine which of the two operands was the base register and which was the index. So we can determine diff --git a/gcc/testsuite/gcc.dg/torture/pr113255.c b/gcc/testsuite/gcc.dg/torture/pr113255.c new file mode 100644 index 0000000..2f00952 --- /dev/null +++ b/gcc/testsuite/gcc.dg/torture/pr113255.c @@ -0,0 +1,27 @@ +/* { dg-do run } */ +/* { dg-additional-options "-mtune=k8 -mstringop-strategy=rep_8byte" { target { x86_64-*-* i?86-*-* } } } */ + +struct S { unsigned a[10]; unsigned y; unsigned b[6]; } g[2]; + +__attribute__((noinline, noclone)) int +test (int x) +{ + struct S e[2] = { g[0], g[1] }; + int r = 0; + if (x >= 0) + { + r++; + e[1].y++; + } + g[1] = e[1]; + return r; +} + +int +main () +{ + test (1); + if (g[1].y != 1) + __builtin_abort (); + return 0; +} |