11:51 <kcxt> xypron: I noticed you print the frame pointer address in your backtrace patches, is that useful for debugging?

11:52 <xypron> kcxt: The suggested patch arm64 needs rework. But I did not find time for it.

11:56 <kcxt> xypron: all good, I rebased my symbol lookup patches on top of your series, planning to post them today. if you're happy with them you could pick them and include them in your respin maybe?

11:57 <xypron> kcxt: Just pick from my patch what you need. If you don't want the framepointer address, remove it.

11:58 <xypron> kcxt: Best send a single series with all the required patches including mine.

11:59 <kcxt> ah ok, sure thing

12:09 <kcxt> xypron: so it's fine if i just send this as v2 of your series with symbol support added?

12:12 <xypron> kcxt: Yes, please.

13:03 <kcxt> xypron: oki, patches sent! please give them a whirl

13:03 <kcxt> just enable CONFIG_SYMBOL_LOOKUP

13:04 <kcxt> i made it select FRAMEPOINTER but in hindsight that isn't really necessary, they're quite independent

19:59 <Tartarus> Well, here's a fun one I'll post on the ML soon'ish. With LTO, and a big enough U-Boot binary (in this case am64x_evm_a52_defconfig + UNIT_TESTS - UT_OPTEE + some EFI tests too) we get some "relocation truncated to fit" link problems, unless I move arch/arm/cpu/armv8/exceptions.o and arch/arm/lib/relocate_64.o to be next to start.o. I don't like that, and furthermore I'm only right now trying to boot test this change on both the

19:59 <Tartarus> failing to link platform and a few other armv8 platforms.