ChanServ changed the topic of #linux-sunxi to: Allwinner/sunxi development - Did you try looking at our wiki? https://linux-sunxi.org - Don't ask to ask. Just ask and wait for an answer! - This channel is logged at https://oftc.irclog.whitequark.org/linux-sunxi
gamiee has quit [Ping timeout: 480 seconds]
apritzel has quit [Ping timeout: 480 seconds]
gnarface has quit [Server closed connection]
gnarface has joined #linux-sunxi
palmer has quit [Server closed connection]
cnxsoft has joined #linux-sunxi
palmer has joined #linux-sunxi
Asara has quit [Server closed connection]
Asara has joined #linux-sunxi
ftg has quit [Ping timeout: 480 seconds]
aggi has quit [Quit: leaving]
macromorgan has quit [Server closed connection]
macromorgan has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
sh1 has quit [Ping timeout: 480 seconds]
sh1 has joined #linux-sunxi
adjtm has quit [Server closed connection]
adjtm has joined #linux-sunxi
sunshavi has quit [Server closed connection]
sunshavi has joined #linux-sunxi
rajkosto has joined #linux-sunxi
rajkosto has quit [Quit: Leaving]
gamiee has joined #linux-sunxi
apritzel has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
libv_ is now known as libv
apritzel has joined #linux-sunxi
cnxsoft has quit [Ping timeout: 480 seconds]
sunshavi has quit [Remote host closed the connection]
sunshavi has joined #linux-sunxi
evgeny_boger has joined #linux-sunxi
evgeny_boger has quit []
veremitz has quit [Quit: ZNC - http://znc.in]
veremitz has joined #linux-sunxi
evgeny_boger has joined #linux-sunxi
suniel has joined #linux-sunxi
<maz> aperezdc: Amlogic went out of their way to fit a GIC400 there.
<maz> apritzel: ^^
<maz> they just disabled the v3 interface and routed everything out. pretty unexpected, but still functional.
<apritzel> maz: I see, thanks. I was under the impression that there would be something missing for that in the A55 core
veremitz has quit [Read error: Connection reset by peer]
<apritzel> so they effectively treat the GIC like any other bog standard MMIO device?
<maz> apritzel: they just drive the 4 IRQ/FIQ pins externally (no stream interface), and yes, this is just MMIO. only worse due to the extra latency.
veremitz has joined #linux-sunxi
<apritzel> oh, I see ...
<maz> but hey, my gateway VM runs on that, and it's been happy for a couple of years.
<apritzel> I guess AW is now disappointed that nobody told them about this and they licensed a newer GIC ;-)
<apritzel> (because those kinds of hacks sound exactly like what AW would do to save yet another penny)
<maz> don't worry, they'll fsck it up another ywa.
<apritzel> oh, yes, I am pretty sure about that!
<maz> is the HW available? or is that just slideware?
<maz> I need a new GICv3 box with 8.2 CPUs (without buying an Ampere box).
<apritzel> apparently code snippets appeared in some BSP drop (for another SoC)
<maz> right. probably another 12/18 months before it actually appears then.
<apritzel> exactly
<apritzel> maz: those RK3568 boards should tick the box for you?
<maz> apritzel: yeah, to some extent. they borked the ITS and RDs though (the attributes are not probable, and they can only access the bottom 32bit).
<apritzel> ah, of course ;-)
<maz> the workarounds they submitted upstream were a total mess, so they are simply not used for MSIs upstream.
<maz> it I get a board one day, I'll try and fix it.
<maz> (well, fix is a strong word... paper over it more likely)
sh1 has quit [Ping timeout: 480 seconds]
adjtm is now known as Guest771
adjtm has joined #linux-sunxi
cnxsoft has joined #linux-sunxi
Guest771 has quit [Ping timeout: 480 seconds]
JohnDoe_71Rus has quit []
evgeny_boger has quit [Ping timeout: 480 seconds]
suniel has quit [Remote host closed the connection]
ftg has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
cnxsoft has quit []
veremitz has quit [Quit: ZNC - http://znc.in]
veremitz has joined #linux-sunxi
v3r3mitz has joined #linux-sunxi
veremitz has quit [Ping timeout: 480 seconds]
v3r3mitz is now known as veremitz
veremitz has quit [Quit: ZNC - http://znc.in]
adjtm has quit [Quit: Leaving]
veremitz has joined #linux-sunxi
veremitz has quit []
suniel has joined #linux-sunxi
veremitz has joined #linux-sunxi
<jernej> apritzel: what is biggest improvement in GICv3? I mean, what incentive would company have to integrate newer GIC, if old will do?
<apritzel> jernej: two killer features: support for more than 8 cores (this is a hard architectural limit in GICv2), plus using system registers for the GIC CPU interface instead of MMIO
<apritzel> per-core banked MMIO has higher latency and is harder to implement in modern SoCs
<jernej> so 1) not relevant for AW, at least with existing SoCs
<apritzel> exactly
<jernej> and second?
<apritzel> second is the system register interface I mentioned, which is architectural and integrated into the core
<jernej> so lower latency as you mentioned before?
<jernej> is this latency really so much relevant for non RT work?
<apritzel> no
<apritzel> also: newer Arm Ltd. cores (starting with the ARMv8.2 parts, so A55/A75 and beyond) don't support the GICv2 CPU side anymore, at least on the face of it
<apritzel> that is probably more of an issue, but as Amlogic proved, it can be hack^Wworked around
<jernej> ok, this sounds more important for me :)
<jernej> ah, many things can be achieved, given enough resources and motivation
<apritzel> so that's what I meant on the weekend: there is really no *pressing* reason for AW to use a GICv3, and they strike me as the kind of vendor pulling of this GIC-400 hack that Amlogic did
<gamiee> well, Allwinner is recently quite silent, almost no new (ARM) SoCs out (only D1 and D1s, but that's... at least for me not that interesting)
<apritzel> MSIs are handled better with a GICv3 (admittedly at the cost of quite some complexity with the ITS), but for that PCIe level that AW uses it doesn't seem relevant
<apritzel> gamiee: I think they released quite some SoCs, but is mostly specialised (read: not-interesting) SoCs, many with Cortex-A7 still
<gamiee> apritzel: which ones?
<apritzel> gamiee: various in the R and T series
<gamiee> will take look, thx
<apritzel> gamiee: some are not really well advertised, but there is evidence in some BSP code. It smells like very specific chips for one use case (for instance dashboard cameras or "smart" speakers)
evgeny_boger has joined #linux-sunxi
evgeny_boger has quit [Ping timeout: 480 seconds]
evgeny_boger has joined #linux-sunxi
vagrantc has joined #linux-sunxi
hlauer has joined #linux-sunxi
evgeny_boger has quit [Ping timeout: 480 seconds]
apritzel has quit [Ping timeout: 480 seconds]
vagrantc has quit [Quit: leaving]
JohnDoe_71Rus has quit []
apritzel has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
<gamiee> apritzel: yes, V-series chips are interesting, but I don't understand why Allwinner hiding their V831/V833 that much (also why they hides the fact it haves NNA)
Luke-Jr has quit [Ping timeout: 480 seconds]
Luke-Jr has joined #linux-sunxi
vagrantc has joined #linux-sunxi
suniel has quit [Remote host closed the connection]
evgeny_boger has joined #linux-sunxi
apritzel has joined #linux-sunxi
<jernej> gamiee: any luck with CEC wakeup?
Luke-Jr has quit [Ping timeout: 480 seconds]
hlauer has quit [Ping timeout: 480 seconds]
Luke-Jr has joined #linux-sunxi
samueldr has quit [Server closed connection]
samueldr has joined #linux-sunxi
eldondev has quit [Server closed connection]
eldondev has joined #linux-sunxi
sh1 has joined #linux-sunxi