| *** tpb <[email protected]> has joined #openrisc | 00:00 | |
| shorne | zx2c4: kernel patches now here: https://github.com/stffrdhrn/linux/commits/or1k-virt | 00:12 |
|---|---|---|
| shorne | I still get hangs during bootup sometimes, like 2-3 seconds to go from kicking /init to userland | 00:13 |
| shorne | sometimes it seems to never get to userland | 00:13 |
| shorne | I'll keep doing testing | 00:14 |
| shorne | first needs to test my PIC change on the "real" hardware | 00:43 |
| shorne | ok, it works on the litex soc fpga board | 01:04 |
| shorne | ah, but thats not using level triggered pic driver | 01:06 |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 03:09 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 03:14 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 03:23 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 03:25 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 05:42 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 05:59 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 06:12 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 06:13 | |
| *** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 07:12 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 07:16 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 07:18 | |
| *** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 07:21 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 07:25 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 07:28 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 08:34 | |
| zx2c4 | shorne: alright will give it a try | 08:48 |
| zx2c4 | oh, looks like you pushed a pic commit? | 08:49 |
| shorne | zx2c4: yes for the kernel | 08:51 |
| shorne | the pic level driver has and issue that manifests in SMP systems | 08:51 |
| zx2c4 | oh yea im seeing huge stalls still :-\ | 08:52 |
| zx2c4 | hrtimer related i think? | 08:52 |
| shorne | basically, the interrupt controller (in qemu and most soft cores) is level triggered meansing the only way to clear the interrupt is by interacting with the hardware | 08:53 |
| shorne | hmm, what kind of huge stalls? | 08:53 |
| shorne | during boot time? or after starting your CI tests? | 08:54 |
| zx2c4 | https://א.cc/7DM7ERD4 | 08:54 |
| zx2c4 | when starting tests, during tests. sometimes they resolve after 3-15 seconds. othertimes i get this splat | 08:55 |
| shorne | right | 08:55 |
| shorne | ok, I may have seen that | 08:55 |
| shorne | I'll investigate, need to instrument the kernel | 08:55 |
| shorne | its with accel=tcg=multi and how many cpus? | 08:56 |
| shorne | i see ncpus=4 | 08:56 |
| zx2c4 | 4, yea | 08:56 |
| shorne | ok, not quiet there yet | 08:57 |
| shorne | there might be other irq related issues, fixing the level pic in the kernel helped a bit, but I think there might be an underlying issue in qemu (openrisc implementation) | 08:58 |
| shorne | it could be timers too as you mentioned | 08:58 |
| shorne | though I ahvent seen the 200s jumps today | 08:58 |
| shorne | got to go.. pick up kids | 08:58 |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 09:26 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 09:37 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 09:48 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 14:22 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 15:24 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 17:24 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 18:11 | |
| *** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 18:39 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 18:42 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 18:44 | |
| *** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 18:45 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 19:44 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 19:50 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 19:57 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 19:59 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Remote host closed the connection) | 20:16 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 20:16 | |
| shorne | zx2c4: investigating the slowness more, I have looked into this before (a few years back when running glibc tests on qemu) | 21:34 |
| shorne | the issue before was the openrisc tlb is highly constrained and qemu ends up spending so much time doing tlb ping-pong | 21:34 |
| shorne | it looks similar now, I am looking to see if I can try to put in some optimizations (i.e. multi way tlb) to see if it helps | 21:35 |
| shorne | before for glibc testing I ended up just moving over to fpga hardware which didn't have such issues | 21:35 |
| *** arnd_ <[email protected]> has joined #openrisc | 22:04 | |
| *** arnd <[email protected]> has quit IRC (Ping timeout: 250 seconds) | 22:07 | |
| *** arnd_ is now known as arnd | 22:07 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Remote host closed the connection) | 22:08 | |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc | 22:09 | |
| zx2c4 | shorne: oh you think the hangs and that splat are just the result of tlb slowness on smp4 rather than a deadlock or something? | 22:22 |
| zx2c4 | As usual, walking home from another jazz thing at this hour. So when im back, happy to try out whatever | 22:23 |
| shorne | zx2c4: I think it might be related, I could be wrong, but defintely some place that there is slowness so tuning that will be a plus | 23:12 |
| *** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds) | 23:41 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!