[Bug target/83466] Wrong TLS GD sequence for ILP32
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83466 --- Comment #8 from John Dong --- Created attachment 51045 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51045&action=edit patch to fix pr83466 patch to fix this issue for SYMBOL_SMALL_TLSDESC and SYMBOL_SMALL_TLSIE.
[Bug target/83466] Wrong TLS GD sequence for ILP32
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83466 John Dong changed: What|Removed |Added CC||dongjianqiang2 at huawei dot com --- Comment #7 from John Dong --- (In reply to Steve Ellcey from comment #4) > Created attachment 43027 [details] > Patch file being tested > > I am testing this patch for regressions, I have verified that it does fix > the small test case. can u please (In reply to Steve Ellcey from comment #4) > Created attachment 43027 [details] > Patch file being tested > > I am testing this patch for regressions, I have verified that it does fix > the small test case. can u please fix on trunk ?
[Bug target/83466] Wrong TLS GD sequence for ILP32
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83466 --- Comment #6 from Andreas Schwab --- The same issue exists for SYMBOL_SMALL_TLSIE.
[Bug target/83466] Wrong TLS GD sequence for ILP32
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83466 --- Comment #5 from Andreas Schwab --- This also fixes the original failure: miscompiled efl that causes python3-efl to fail to build.
[Bug target/83466] Wrong TLS GD sequence for ILP32
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83466 --- Comment #4 from Steve Ellcey --- Created attachment 43027 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=43027&action=edit Patch file being tested I am testing this patch for regressions, I have verified that it does fix the small test case.
[Bug target/83466] Wrong TLS GD sequence for ILP32
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83466 Andrew Pinski changed: What|Removed |Added Status|UNCONFIRMED |NEW Last reconfirmed||2017-12-18 Ever confirmed|0 |1 --- Comment #3 from Andrew Pinski --- Yes that should be a 32bit zero extending add due to the ABI.
[Bug target/83466] Wrong TLS GD sequence for ILP32
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83466 --- Comment #2 from Andreas Schwab --- Pasted the wrong sample: adrpx0, :tlsdesc:ti ldr w1, [x0, #:tlsdesc_lo12:ti] add w0, w0, :tlsdesc_lo12:ti .tlsdesccallti blr x1 mrs x1, tpidr_el0 add x0, x1, x0
[Bug target/83466] Wrong TLS GD sequence for ILP32
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83466 --- Comment #1 from Andrew Pinski --- >ldr x1, [x0, #:tlsdesc_lo12:ti] This most likely should be a 32bit load also.