Monday, 2022-05-09

*** tpb <[email protected]> has joined #openrisc00:00
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc00:21
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)00:57
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc01:11
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)01:58
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc02:40
shorneok, one more thing I thought of, we have another simulator or1ksim, I tried booting the faulty kernel on that, I cannot reproduce the issue there either03:19
shorneso more things pointing to qemu03:19
*** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc04:30
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)04:31
*** shorne <[email protected]> has quit IRC (*.net *.split)05:11
*** shorne <[email protected]> has joined #openrisc05:11
*** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)05:16
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc05:22
*** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc05:27
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)05:30
*** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)05:31
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc05:32
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)05:36
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc05:42
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)07:51
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc07:52
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)08:13
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc08:15
*** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc08:44
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)08:46
zx2c4shorne: i wonder if do_range() is doing something silly with the branch delay slot?10:38
zx2c4Or openrisc_cpu_do_interrupt10:39
shornezx2c4: yeah, it could be, I have been reading the code to try to remember it better10:47
shorneI bet it is something to do with the delay slot handling10:59
zx2c4shorne: which would correspond with the branches in multdi310:59
zx2c4Missing some subtle carry10:59
zx2c4Making crypto selftests go wild10:59
zx2c4Whereas most arithmetic doesnt really care about overflow bit10:59
shornebut it seems to be some combintation with the alignment too11:01
shornethere is also a condition where translation blocks are exited due to TOO_MANY instructions (DISAS_TOO_MANY), I am wondering if there is a problem when that falls on a delay slot boundary11:44
*** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)11:47
shornemaybe not though because the disasemble context that tracks delay slot doesnt reset for new blocks11:56
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc11:56
shornewell, it tries to remember the delayslot11:58
zx2c4hmm12:07
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)13:38
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc13:39
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)13:47
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc15:13
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)15:45
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc20:29
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)21:36
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc21:38
*** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc21:57
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)21:59
*** littlebo1eep <littlebo1eep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)22:10
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc22:17
shorneone more clue, if we look at the disassembly of fe_mul_impl22:34
shornehttps://gist.github.com/stffrdhrn/32792410a3df5d0f1534a1387765ff7c22:34
shorneIt falls on a page boundary, and that is a delay slot22:35
shornec01e3ffc:10 00 00 03 l.bf c01e4008 <fe_mul_impl+0x9ac>22:35
shorneas I understand, qemu splits basic blocks on page boundaries22:36
zx2c4shorne: bingo!23:05
zx2c4so maybe it's not considering the delay as part of the block23:05
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)23:20
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc23:22
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)23:27
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has joined #openrisc23:32
*** littlebobeep <littlebobeep!~alMalsamo@gateway/tor-sasl/almalsamo> has quit IRC (Ping timeout: 240 seconds)23:55

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!