*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC | 05:47 | |
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc | 06:01 | |
*** Guest27 <Guest27!~Guest27@151.70.146.154> has joined #libre-soc | 08:26 | |
*** Guest27 <Guest27!~Guest27@151.70.146.154> has quit IRC | 08:58 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.54.52> has joined #libre-soc | 09:37 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.54.52> has quit IRC | 10:40 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.172.1> has joined #libre-soc | 10:41 | |
lkcl | from pabs on #debian-ppc: https://www.talospace.com/2023/07/firefox-115-on-power.html | 12:00 |
---|---|---|
lkcl | sadoon[m], toshywoshy, markos_ ^ | 12:00 |
markos_ | yeah saw that already but using my talos headless for a while | 12:06 |
sadoon[m] | Unfortunately it's way behind chromium which works perfectly with JIT as well | 12:27 |
sadoon[m] | One of the reasons I've left firefox behind for a while | 12:27 |
sadoon[m] | Not having to build my browser at every update is a huge plus | 12:28 |
sadoon[m] | Classilla used to thanklessly build his own JIT patch for firefox back in 91esr days but it's been broken since | 12:28 |
sadoon[m] | That's the last time I used firefox :p | 12:29 |
lkcl | sadoon[m], joooy | 13:29 |
lkcl | markos_, hey what's wrong with xdrp and rdesktop? :) | 13:29 |
toshywoshy | yes, FF has been able to build for about 3 months now, but I haven't had the time to replicate it yet | 13:34 |
lkcl | folks i need all of you to help review the Schedule A here https://bugs.libre-soc.org/show_bug.cgi?id=961 | 13:40 |
lkcl | ghostmansd[m], cesar, toshywoshy, programmerjake, markos_, sadoon[m], shriyasharma[m] ^ | 13:40 |
lkcl | each bugreport of https://bugs.libre-soc.org/show_bug.cgi?id=961#c0 has to match the corresponding bugreport for both description, title, and EUR amount. | 13:41 |
lkcl | this needs to be done expediently because there are some tasks completed and payments are held up until the MoU has been signed | 13:42 |
*** ghostmansd <ghostmansd!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has joined #libre-soc | 13:42 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.172.1> has quit IRC | 13:43 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has joined #libre-soc | 13:44 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has quit IRC | 14:22 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.172.116> has joined #libre-soc | 14:24 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.172.116> has quit IRC | 14:44 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@85.30.228.54> has joined #libre-soc | 14:45 | |
*** Guest27 <Guest27!~Guest27@151.70.146.154> has joined #libre-soc | 15:28 | |
sadoon[m] | Just have some laundry to get sorted and then I'll look into it all | 16:11 |
*** Guest27 <Guest27!~Guest27@151.70.146.154> has quit IRC | 16:17 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@85.30.228.54> has quit IRC | 16:21 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.162.28> has joined #libre-soc | 16:22 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.162.28> has quit IRC | 16:32 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.53.227> has joined #libre-soc | 16:35 | |
sadoon[m] | Also I changed my email recently to a new domain, is there a way to change that on bugzilla? | 16:46 |
sadoon[m] | I looked into task 999 and fixed any inconsistencies, now everything matches section A | 16:53 |
sadoon[m] | <lkcl> "rats, put it on the list https:/..." <- FrOSCON is already on that list, not sure if you added it recently though | 16:55 |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.53.227> has quit IRC | 16:59 | |
sadoon[m] | lkcl you mentioned on section A bug 999 that the Linux kernel already has emulation of v3.0 even on older ISAs | 17:01 |
sadoon[m] | It just hit me, if I emulate POWER7, which has no VSX in little endian, I could get closer to the goal of verifying VSX isn't in our code | 17:01 |
sadoon[m] | In fact I don't even think it has altivec there on LE does it? | 17:03 |
sadoon[m] | would be huge if it worked | 17:03 |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.168.124> has joined #libre-soc | 17:04 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.168.124> has quit IRC | 17:12 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.168.124> has joined #libre-soc | 17:13 | |
sadoon[m] | And this is my aha moment | 17:27 |
sadoon[m] | Kernel panic on power7 but not power9 | 17:27 |
sadoon[m] | make -j72 a kernel for power7 coming up next | 17:28 |
sadoon[m] | (packages still built for "sffs", just the kernel for power7 and altivec and vsx disabled kernel wide) | 17:28 |
programmerjake | i think the budget on https://bugs.libre-soc.org/show_bug.cgi?id=1026 needs to be bigger, eur 8000 could be justified for fptrans by itself... | 17:37 |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.168.124> has quit IRC | 17:39 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has joined #libre-soc | 17:40 | |
programmerjake | other than that looks fine, though I only checked #1026 and didn't check if all the other bugs match | 17:42 |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has quit IRC | 18:01 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.168.159> has joined #libre-soc | 18:01 | |
sadoon[m] | How am I getting an illegal instruction on power9 AND power7 | 18:12 |
programmerjake | which opcode is the illegal instruction? | 18:13 |
sadoon[m] | modprobe[50]: illegal instruction (4) at 7fff9036ffd0 nip 7fff9036ffd0 lr 7fff90363bfc code 1 in ld64.so.2[7fff90330000+50000] | 18:14 |
sadoon[m] | illegal instruction (4) at 7fffbe67ffd0 nip 7fffbe67ffd0 lr 7fffbe673bfc code 1 in ld64.so.2[7fffbe640000+50000] | 18:14 |
programmerjake | hmm, weird that you're getting an illegal insn...it doesn't say what the opcode is... | 18:17 |
sadoon[m] | This looks kindof like what would happen if I didn't rebuild some software for sffs while other stuff is built on that..? | 18:17 |
programmerjake | maybe it sees that the cpu is POWER7 so doesn't bother to check for supported features? | 18:19 |
sadoon[m] | But then it should run fine on power9 | 18:19 |
programmerjake | or, did you forget to rebuild ld.so? | 18:19 |
sadoon[m] | programmerjake: I tried to rebuild everything, but perhaps there's different "everything"s for gentoo | 18:19 |
sadoon[m] | Might have skipped some | 18:19 |
sadoon[m] | Let me check | 18:19 |
programmerjake | but not power9 with disabled features... | 18:19 |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.168.159> has quit IRC | 18:20 | |
programmerjake | try attaching gdb to qemu (outside of the vm, not run gdb inside) and see if you can disassemble where the illegal insn is | 18:21 |
programmerjake | and/or see where the source code is | 18:21 |
programmerjake | i guess it could be from some assembly file that sees #ifdef POWER7 and runs some op you disabled in qemu | 18:22 |
sadoon[m] | I haven't disabled anything just yet, only VSX in kconfig | 18:23 |
sadoon[m] | Only the kernel crashes, running a chroot is fine | 18:23 |
programmerjake | well, that's disabling something... | 18:23 |
sadoon[m] | Yeah I guess | 18:23 |
sadoon[m] | I'll try not disabling anything | 18:23 |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has joined #libre-soc | 18:24 | |
sadoon[m] | Trying the gentoo iso | 18:25 |
sadoon[m] | to be absolutely certain | 18:25 |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has quit IRC | 18:29 | |
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has joined #libre-soc | 18:29 | |
lkcl | sadoon[m], use objdump with disasm and track that address | 19:28 |
lkcl | sadoon[m], you can still log in - the email address is the username. | 19:29 |
lkcl | toshywoshy, how did you get qemu with sffs-only and where is it documented? sadoon is now trying to replicate it and spending an inordinate amount of time | 19:30 |
programmerjake | if it uses ASLR, the address won't correspond to the objdump addresses, hence why I recommended attaching gdb to qemu | 19:30 |
*** ghostmansd <ghostmansd!~ghostmans@broadband-90-154-80-239.ip.moscow.rt.ru> has quit IRC | 20:27 | |
*** Guest27 <Guest27!~Guest27@151.70.146.154> has joined #libre-soc | 21:07 | |
*** Guest27 <Guest27!~Guest27@151.70.146.154> has quit IRC | 21:25 | |
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC | 23:39 | |
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc | 23:49 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!