Re: [PATCH v2] gdb/record: avoid rax usage in i386-avx-record.exp
Guinevere Larsen <[email protected]>
| Newsgroups | gmane.comp.gdb.patches |
|---|---|
| Message-ID | <[email protected]> |
On 8/13/26 11:51 AM, Simon Marchi wrote: > On 8/13/26 9:10 AM, Guinevere Larsen wrote: >> My recent commit, 1c04308f0958cbac83bad22703758139ea6969b6: >> >> Author: Guinevere Larsen <[email protected]> >> Date: Tue Jun 16 16:03:43 2026 -0300 >> gdb/record: add support for BMI2 instructions >> >> Introduced a few tests that rely on the rax register being set before >> entering the function. This isn't reliable, as programs compiled with >> gcc-14 and earlier will have the register. To avoid compiler versions > This sentence ends abruptly. > > It's not clear to me what the problem is, could you explain it in simple > words :)? > > Simon > oops, sorry, need to do more proofreading I guess. The sentence should be "will have the registered clobbered, saving 0 on it" I haven't looked at the disassembly to know why, but something about how GCC 14 encodes the function calls makes EAX be 0 every time, and the test expects eax to be something else. Moving to a different register sidesteps this -- Cheers, Guinevere Larsen it/its she/her (deprecated)