>>>>> "Jakub" == Jakub Jelinek <ja...@redhat.com> writes:
Jakub> I admit I haven't looked at the GDB changes, but how will the debugger know Jakub> if it is an emutls emulation or normal ELF TLS, if the same op is used Jakub> in both cases? Because gdb never implemented DW_OP_form_tls_address, that emultls case has never worked. My view is that, if it ought to work, then this has to be some sort of platform-specific thing. gdb already lets different platforms implement TLS lookup differently; it can be done by the target or by the arch, see target_translate_tls_address. My suspicion is that the history here is just that DW_OP_GNU_push_tls_address was introduced to solve this problem, then formalized in DWARF, and then for some reason never followed up on. Jakub> Also, as this effectively requires the latest unreleased GDB under the Jakub> default options for something that has been working previously, I wonder Jakub> if it e.g. for some time shouldn't be guarded with dwarf_version >= 5 Jakub> (which will require substantial changes in gdb anyway), and only be changed Jakub> back to dwarf_version >= 3 in 2 years or so when newer debugger will be much Jakub> more common. I'd be amenable to that. It did occur to me after sending that the DWARF 5 work is going to require a new gdb anyhow. Tom
