mirror/glibc
2
0
Fork 0
mirror of git://sourceware.org/git/glibc.git synced 2025-04-18 14:30:43 +08:00
Code Issues Packages Projects Releases Wiki Activity

Ensure calculations happen with desired rounding mode in y1lf128

Browse Source 
math/test-float128-y1 fails on x86_64 and ppc64el with gcc 12 and -O3,
because code inside a block guarded by SET_RESTORE_ROUNDL is being moved
after the rounding mode has been restored. Use math_force_eval to
prevent this (and insert some math_opt_barrier calls to prevent code
from being moved before the rounding mode is set).

Fixes #29463

Reviewed-By: Wilco Dijkstra <Wilco.Dijkstra@arm.com>
(cherry picked from commit 2b274fd8c9c776cf70fcdb8356e678ada522a7b0)
This commit is contained in:
Michael Hudson-Doyle 2022-08-12 11:29:31 +12:00 committed by Aurelien Jarno
parent 2bd815d834
commit 3e27919274
2 changed files with 4 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
NEWS
View File

@ -24,6 +24,7 @@ The following bugs are resolved with this release:
[29446] _dlopen now ignores dl_caller argument in static mode
[29485] Linux: Terminate subprocess on late failure in tst-pidfd
[29490] alpha: New __brk_call implementation is broken
[29463] math/test-float128-y1 fails on x86_64
[29528] elf: Call __libc_early_init for reused namespaces
[29537] libc: [2.34 regression]: Alignment issue on m68k when using
[29539] libc: LD_TRACE_LOADED_OBJECTS changed how vDSO library are

3
sysdeps/ieee754/ldbl-128/e_j1l.c
View File

@ -869,10 +869,13 @@ __ieee754_y1l (_Float128 x)
{
/* 0 <= x <= 2 */
SET_RESTORE_ROUNDL (FE_TONEAREST);
xx = math_opt_barrier (xx);
x = math_opt_barrier (x);
z = xx * xx;
p = xx * neval (z, Y0_2N, NY0_2N) / deval (z, Y0_2D, NY0_2D);
p = -TWOOPI / xx + p;
p = TWOOPI * __ieee754_logl (x) * __ieee754_j1l (x) + p;
math_force_eval (p);
return p;
}