Closed stffrdhrn closed 4 years ago
The math
failures were mostly from outdated ulps
. I ran make regen-ulps
and now there is 1 failure:
# failures
FAIL math/check-abi-libm
# test finish: 2020-04-29T01:23:18+00:00
# test time(m): 0
# test for file: math/subdir-tests.sum
# test wrapper: /home/shorne/work/openrisc/or1k-utils/glibc/qemu-or1k-libc-chroot
# summary
86 FAIL (Explained by underflow issue)
848 PASS
3 UNSUPPORTED
The failure API check shows an easy to patch issue, now no unexplained math
issues:
at math/check-abi-libm.out
--- ../sysdeps/unix/sysv/linux/or1k/libm.abilist 2020-04-26 17:18:51.647752923 +0900
+++ /home/shorne/work/gnu-toolchain/build-glibc/math/libm.symlist 2020-04-29 06:45:46.173185212 +0900
@@ -770,0 +771,12 @@ GLIBC_2.29 pow F
+GLIBC_2.31 totalorder F
+GLIBC_2.31 totalorderf F
+GLIBC_2.31 totalorderf32 F
+GLIBC_2.31 totalorderf32x F
+GLIBC_2.31 totalorderf64 F
+GLIBC_2.31 totalorderl F
+GLIBC_2.31 totalordermag F
+GLIBC_2.31 totalordermagf F
+GLIBC_2.31 totalordermagf32 F
+GLIBC_2.31 totalordermagf32x F
+GLIBC_2.31 totalordermagf64 F
+GLIBC_2.31 totalordermagl F
ALL RESOLVED
After updating a few check-abi-*
issues the report is was follows:
After fixing that, there was an issue with the __old_semctl function still being compiled and causing a compile failure due to missing __NR_ipc
. I found we needed to define a shlib-version
file to force the version compatibility. This required recompiling and there is now a new set of failures.
FAIL math/test-double-acos
FAIL math/test-double-asin
FAIL math/test-float32x-acos
FAIL math/test-float32x-asin
FAIL math/test-float64-acos
FAIL math/test-float64-asin
FAIL math/test-ldouble-acos
FAIL math/test-ldouble-asin
/bin/cat
Version 1 pushed upstream. closing this for now.
This follows issue #6 now that we have sorted out the test failures we can rebase and run all tests again. We should try to rebase often until all tests are done and we are ready for upstreaming.
This issue is to keep notes and track progress.
For upstreaming we need to:
binutils-gdb
fixes see or1k-glibc branchgcc
fixes see or1k-glibc branchlinux
fixes Not yet pushed but one fix for clone syscall argument clobberingqemu
fixes (need spec update) to allow user mode FPCR updatingGet all the glibc tests to pass again: