18:13 karolherbst: gnurou: I didn't manage to make the pmu use the nouveau image after secure boot, I've tried that reset thing, but I think I failed
18:13 karolherbst: any ideas?
21:59 karolherbst: imirkin: by the way, is there anything bad about the two dual issue patches (without the pass)? Those are fine by themselves, even if the effect is rather slim
22:03 imirkin_: karolherbst: remind me the patch
22:03 karolherbst: https://github.com/karolherbst/mesa/commit/1e5dfda608fbe7b8a34ea391ad7897d65ddd84c5
22:03 karolherbst: https://github.com/karolherbst/mesa/commit/fc33bbe7b21cc9456ccf1ebfcf0022a0a7c01447
22:03 karolherbst: those
22:05 imirkin_: karolherbst: sgtm
22:05 imirkin_: i can't push them right now, but i'll try to remember tonight
22:05 karolherbst: aweomse :)
22:07 imirkin_: karolherbst: why does dependsOn check for DefDef?
22:08 karolherbst: good question actually. I am sure it won't work for dual issue though, even if defSrc should already catch any possible situation
22:08 karolherbst: but well
22:08 imirkin_: ok, so i just don't like the name
22:08 karolherbst: mul $r1 $r2 $r3; mul $r1 $r4 $r5 shouldn't be dual issueable
22:08 karolherbst: even though the first $r1 doesn't matter
22:09 imirkin_: i guess dependsOn is OK
22:09 imirkin_: i'll add more comments
22:10 karolherbst: okay
22:10 imirkin_: to make it obvious that it's actually stronger than "depends on"
22:10 karolherbst: I see
22:10 karolherbst: well
22:10 imirkin_: but also "does not screw with"
22:10 imirkin_: :)
22:10 karolherbst: :D
22:12 karolherbst: I think over the next days I will post my most trivial opt patches as well
22:12 karolherbst: stuff like "max(abs(a), 0) to a" and "sub(a, 0) to a"
22:13 karolherbst: and I should really improve that one: https://github.com/karolherbst/mesa/commit/367936aeac8811d5da1eb639a72467165feead05
22:14 imirkin_: i'd kinda rather you declared war on OP_SUB and removed all usage of it (although actually removing the op will be a pain for various reasons)
22:14 karolherbst: right
22:15 karolherbst: OP_SUB is really ugly to deal with :/
22:15 karolherbst: there is no sub instruction on nvidia hardware anyway, right?
22:15 imirkin_: right
22:15 karolherbst: yeah, we should remove it
22:15 imirkin_: it's just nice to do something like the algebraic sub(a, add(a, b)) == b
22:15 imirkin_: er, i even got that wrong, but you get the idea
22:16 karolherbst: well, neg pases will handle that
22:16 imirkin_: i don't know that they will.
22:16 imirkin_: but they could :)
22:16 karolherbst: mhh maybe you are right
22:17 karolherbst: sub(a, add(a,b)) == add(a, neg(add(a,b))) = add(a, add(neg(a), neg(b)))
22:17 karolherbst: mhh
22:17 karolherbst: do we have a pass to opt the latter?
22:18 imirkin_: i'm sure we don't
22:18 imirkin_: we don't have a lot of algebraic manipulation
22:18 karolherbst: mhhh
22:18 karolherbst: the opt would do something like that: search for add chains
22:18 karolherbst: and try to match a <-> neg(a) pairs
22:18 karolherbst: and remove those
22:19 karolherbst: mhh, maybe if I find some time I could try to do that
22:19 karolherbst: trello card for that? .D
22:19 imirkin_: non-actionable
22:20 karolherbst: ohh, I also have my "set,and,cvt to set" opt
22:20 karolherbst: no idea why I wrote this one though
22:23 karolherbst: imirkin_: do you happen to know why I wrote this one? https://github.com/karolherbst/mesa/commit/d1b911532260c1032a57bec8bfee9fa72e1e6d06
22:24 imirkin_: i thought i pushed something like that
22:24 karolherbst: mhh
22:24 karolherbst: I will look
22:24 karolherbst: maybe I didn'T wrote it at all
22:25 imirkin_: https://cgit.freedesktop.org/mesa/mesa/tree/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp#n243
22:25 imirkin_: certainly looks similar :p
22:25 karolherbst: yes
22:25 karolherbst: https://cgit.freedesktop.org/mesa/mesa/commit/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp?id=71ad8a173f5c64d6384c13f04361455571c42ffe
22:25 karolherbst: ahhh
22:25 karolherbst: you wrote it
22:26 imirkin_: hakzsam noticed the issue, but wrote a very wrong patch to fix it
22:26 karolherbst: I am wondering why I ran into that issue
22:26 imirkin_: it'd come out as render corruption