sorear changed the topic of #riscv to: RISC-V instruction set architecture | https://riscv.org | Logs: https://libera.irclog.whitequark.org/riscv | Matrix: #riscv:catircservices.org
TMM_ has quit [*.net *.split]
raym has quit [*.net *.split]
Perflosopher has quit [*.net *.split]
balrog has quit [*.net *.split]
Leopold_ has quit [*.net *.split]
Bluefoxicy has quit [*.net *.split]
mwette has quit [*.net *.split]
KombuchaKip has quit [*.net *.split]
gordonDrogon has quit [*.net *.split]
mps has quit [*.net *.split]
mmind00 has quit [*.net *.split]
raym has joined #riscv
gordonDrogon has joined #riscv
terminalpusher has joined #riscv
Leopold has joined #riscv
TMM_ has joined #riscv
balrog has joined #riscv
Bluefoxicy has joined #riscv
mmind00 has joined #riscv
heat has joined #riscv
Perflosopher has joined #riscv
powderhorn has quit [Ping timeout: 246 seconds]
mps has joined #riscv
mwette has joined #riscv
khem has quit [Quit: Connection closed for inactivity]
heat has quit [Ping timeout: 245 seconds]
sakman has joined #riscv
Leopold has quit [Ping timeout: 246 seconds]
Leopold has joined #riscv
Andre_Z has quit [Quit: Leaving.]
freakazoid332 has joined #riscv
frkzoid has quit [Ping timeout: 258 seconds]
stolen has joined #riscv
HumanGeek has quit [Remote host closed the connection]
HumanGeek has joined #riscv
Tenkawa has quit [Quit: Was I really ever here?]
sakman_ has joined #riscv
sakman has quit [Ping timeout: 250 seconds]
vagrantc has quit [Quit: leaving]
Tenkawa has joined #riscv
ntwk has quit [Quit: ntwk]
Tenkawa has quit [Quit: Was I really ever here?]
Bluefoxicy has quit [Ping timeout: 246 seconds]
davidlt has joined #riscv
BootLayer has joined #riscv
terminalpusher has quit [Ping timeout: 246 seconds]
sakman_ is now known as sakman
Bluefoxicy has joined #riscv
khem has joined #riscv
EchelonX has quit [Quit: Leaving]
valdemaras has joined #riscv
valdemaras has quit [Remote host closed the connection]
valdemaras has joined #riscv
zjason` is now known as zjason
crabbedhaloablut has joined #riscv
jacklsw has joined #riscv
MaxGanzII has joined #riscv
valdemaras has quit [Quit: valdemaras]
Wickram has joined #riscv
valdemaras has joined #riscv
mobius has quit [Ping timeout: 246 seconds]
mobius has joined #riscv
davidlt has quit [Ping timeout: 245 seconds]
valdemaras has quit [Quit: valdemaras]
valdemaras has joined #riscv
junaid_ has joined #riscv
davidlt has joined #riscv
Kedleston_ has joined #riscv
Kedleston has quit [Ping timeout: 256 seconds]
Kedleston_ is now known as Kedleston
junaid_ has quit [Remote host closed the connection]
BootLayer has quit [Quit: Leaving]
valdemaras has quit [Quit: valdemaras]
junaid_ has joined #riscv
ntwk has joined #riscv
valdemaras has joined #riscv
BootLayer has joined #riscv
Tenkawa has joined #riscv
davidlt has quit [Ping timeout: 246 seconds]
junaid_ has quit [Remote host closed the connection]
mahk has quit [Ping timeout: 244 seconds]
Tenkawa has quit [Quit: Was I really ever here?]
stolen has quit [Quit: Connection closed for inactivity]
junaid_ has joined #riscv
Wickram has quit [Quit: WeeChat 4.0.2]
<Esmil> mps: is "vendor (starfive/debian) u-uboot" different from this? https://github.com/starfive-tech/VisionFive2/releases/
<mps> Esmil: this is vendor u-boot afaik, I use latest v3.4.5
<Esmil> i see. my board might have an older version that doesn't initialize pci
<Esmil> ..but linux driver should really be doing that too
mahk has joined #riscv
<mps> Esmil: this u-boot works fine and boots kernel from nvme without problem
<mps> mainline u-boot doesn't work
joev1 has quit [Ping timeout: 260 seconds]
joev1 has joined #riscv
davidlt has joined #riscv
MaxGanzII has quit [Ping timeout: 246 seconds]
EchelonX has joined #riscv
Andre_Z has joined #riscv
mahk has quit [Changing host]
mahk has joined #riscv
terminalpusher has joined #riscv
junaid_ has quit [Remote host closed the connection]
sakman has quit [Ping timeout: 255 seconds]
junaid_ has joined #riscv
joev1 has quit [Ping timeout: 255 seconds]
joev1 has joined #riscv
indy has quit [Quit: ZNC 1.8.2 - https://znc.in]
paddymahoney has quit [Ping timeout: 256 seconds]
billchenchina has joined #riscv
<leah2> anyone know how to see which firmware a visionfive2 has flashed/
indy has joined #riscv
<mps> leah2: you mean u-boot and opensbi or something else
Stat_headcrabed has joined #riscv
<leah2> yes
<leah2> https://github.com/starfive-tech/VisionFive2/releases they are here, but i don't know what i'm running currently
jmdaemon has quit [Ping timeout: 246 seconds]
Jackneill has joined #riscv
sakman has joined #riscv
<mps> it should show version for both on console
<leah2> yeah i need to get a serial working :>
<leah2> but i left the adapter at home
<mps> ah, you don't have serial console
<leah2> i thoguht i can run strings on /dev/mtd? but it's nontrivial
Nixkernal has joined #riscv
MaxGanzII has joined #riscv
valdemaras has quit [Quit: valdemaras]
<leah2> ok, new firmware boots from nvme ++
valdemaras has joined #riscv
jacklsw has quit [Ping timeout: 255 seconds]
Tenkawa has joined #riscv
<Tenkawa> Esmil: first round of tests didn't see the nvme drive on boot.. it did start the kernel but no nvme detected.
<Tenkawa> Nothing obvious showed up as to why though
<Esmil> Tenkawa: the config option probably changed name with Minda's latest patches
<Tenkawa> oh?
<Esmil> i have CONFIG_PCIE_STARFIVE_HOST=m
<Tenkawa> let me check that
<Tenkawa> (this is where the multi boot really helps)
<Esmil> You'll alse need the new clock drivers: CONFIG_CLK_STARFIVE_JH7110_PLL=y CONFIG_CLK_STARFIVE_JH7110_SYS=y CONFIG_CLK_STARFIVE_JH7110_AON=m CONFIG_CLK_STARFIVE_JH7110_STG=m CONFIG_CLK_STARFIVE_JH7110_ISP=m CONFIG_CLK_STARFIVE_JH7110_VOUT=m
<Tenkawa> ok the first one is set.. let me check those others
<Esmil> CONFIG_RESET_STARFIVE_JH7110=y
<Esmil> ..and the pcie phy CONFIG_PHY_STARFIVE_JH7110_PCIE=m
<Tenkawa> all set
<Tenkawa> it boots fine
<Tenkawa> just doesn't see the nvme
<Tenkawa> to mount root
<Tenkawa> once the initram handoff occurs
<Esmil> also make sure that those drivers are actually in your initramfs
<Tenkawa> they are hardcoded into the kernel.. would that matter?
<Esmil> No, but something in your end prevents them from initializing pci/nvme
<Tenkawa> yeah I was suspicious pci might the problem
<Esmil> another thing could be if you didn't load the device tree that was built from the same source
<Tenkawa> Yeah thats "very" possible
<Tenkawa> I'm still questioning the load mechanism
<Tenkawa> let me go try to ensure I have a forced load
<Tenkawa> dtb
<Tenkawa> fixed
<Tenkawa> was copied in the wrong spot
<Tenkawa> the install routine missed the second path
<Tenkawa> Linux vf2 6.5.0-rc7-vf2+ #2 SMP Fri Aug 25 21:32:51 EDT 2023 riscv64 GNU/Linux
<Tenkawa> Timing O_DIRECT disk reads: 1006 MB in 3.00 seconds = 334.78 MB/sec
<Tenkawa> not bad
<Tenkawa> No IO errors so far
<Tenkawa> I'm not even using the updated IO tuning boot params
<Tenkawa> I took them out as a test
<Esmil> interesting. i still get the 'nvme0: I/O 640 QID 5 timeout, completion polled' even with v3.4.5
Nixkernal has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
<Stat_headcrabed> Esmil: Btw what's your cache controller series' status for vf2?
<Esmil> ?
<Tenkawa> major improvement here... just did a big io op and flying
<Esmil> Stat_headcrabed: i'm not sure what you mean
<Stat_headcrabed> I remember you have a patch for solving cache coherence problem?
<Stat_headcrabed> Only jh7100 or also work for jh7110?
<Esmil> That's for the JH7100 (vf1)
<Tenkawa> still no messages in kernel ring
<Tenkawa> This is really nice
<Stat_headcrabed> What do we need for solving this problem on jh7110 ?
<Esmil> Stat_headcrabed: the jh7110 doesn't have that problem
<Tenkawa> Temp: 38.3°C using the ice cooler
<Stat_headcrabed> Nope, 7110 has this problem on some peripherals
<Stat_headcrabed> Mostly graphics related stuff
<Esmil> Oh, and they want to solve it the same way as the jh7100?
<Stat_headcrabed> Such as display controller
<Esmil> (by manually flushing the cache through the sifive cache controller)
<Stat_headcrabed> seems so
<Stat_headcrabed> I noticed a hack on downstream driver
<Esmil> Stat_headcrabed: yikes :(
<Tenkawa> Esmil: you and tell whoever involved, great job so far... I'll comtinue to run the multitude of tests I have been
<Tenkawa> er continue
<Stat_headcrabed> To make it work correctly, I think we also need to flush l1 dcache?
<Esmil> Stat_headcrabed: you can see the latest version of the patches in the visionfive branch (mine or the starfive repo)
<Stat_headcrabed> Btw we could get rid of most cache coherence problem by plugging in a PCIE gpu :)
<mps> Tenkawa: iiuc these patches gives better performance?
stolen has joined #riscv
<Tenkawa> mps: much
<Tenkawa> I was able to remove all my manual adjustments too
<Tenkawa> the latency seems to be much lower on the io..
<mps> ok, will test it
<mps> have to pick these patches
<Tenkawa> the dtb is "very" important so don't miss it... that got me
<Tenkawa> pieces of it work without it
<Tenkawa> but not everything
<Tenkawa> my make install missed copying it to the right spot
<Tenkawa> initially
<Tenkawa> I just rsync'ed 2.5gb with no IO errors
billchenchina- has joined #riscv
billchenchina has quit [Read error: Connection reset by peer]
<Tenkawa> Thats a good sign
<mps> no worries, I have alpine APKBUILD for building all needed things
<Tenkawa> About to run a full compile
<mps> but i didn't had errors with nvme, except with mainline u-boot
<mps> and with not so good power supply
<Tenkawa> you need good power
<Tenkawa> its crucial
<mps> this one which come with VF2 is fine
<Tenkawa> mps: to run with nvme actually its not
<mps> I don't have any error with it
<Tenkawa> mps: with vendor software yes.... they didn't spec it to run mainline though
<Tenkawa> There are different voltage and current tolerances
<mps> hm, could this be problem I have with mainline u-boot and nvme
<Tenkawa> We've fought the same thid problem in ARM for years
<Tenkawa> er this
<Tenkawa> Its very possible yes
<mps> I have two apple silicon macbooks, could try with their power supply
aerkiaga has joined #riscv
<Tenkawa> Its definitely strong enough... the PD in it is the potential problematic part
<Tenkawa> This is why I use/advise bench/static power
<Tenkawa> I have 4 bench power supplies just for this purpose
<Tenkawa> nice... -j4 compile running.. still only Temp: 39.9°C
<Tenkawa> This is a great improvement
<Tenkawa> 100% cpu used
<Tenkawa> heh
<mps> no luck, apple silicon power supply doesn't help
paddymahoney has joined #riscv
<Tenkawa> Well that's unfortunate.. do you have any way to get the output online?
paddymahoney has quit [Remote host closed the connection]
valdemaras has quit [Quit: valdemaras]
<mps> output online?
<Tenkawa> the boot failure
Wickram has joined #riscv
<mps> ah, I posted it few days ago https://dev.alpinelinux.org/~mps/riscv64/boot-fail
<Tenkawa> let me taake a look
Wickram has quit [Client Quit]
<Tenkawa> you still didn't try adding those nvme settings did you?
<Tenkawa> with that kernel until you do... you are likely going to continue to keep getting them
<mps> what kernel settings?
<mps> s/what/which/
<Tenkawa> nvme.io_queue_depth=2048 nvme.max_host_mem_size_mb=512 nvme.write_queues=4 nvme.poll_queues=4 nvme_core.default_ps_max_latency_us=0 nvme_core.apst_secondary_timeout_ms=100
<Tenkawa> test those on your boot command line
<mwette> maybe trolling, but do you know about nvme-cli (I found via web search)
<mwette> ^userspace command line interface to nvme
<Tenkawa> mwette: thats great if you want to change "disk" settings.. not kernel settings
<mwette> got it
<Tenkawa> These are settings for how the nvme kernel driver communicates/operates
<mps> Tenkawa: ok, will add and check
<Tenkawa> mwette: I used nvme-cli to change blocksize but changed it back after I learned a "unfortunate" side effect
<Tenkawa> It may slightly help perf but after you get certain sized drives by certain vendors... their part tables are a mess to work wiith
<Tenkawa> And since none of these socs can support x4 speeds yet.. its not really worth changing the mode
<mps> I use nvme-cli only as information tool
<Tenkawa> ooh.. up to 41.8C
<Tenkawa> lol
<Tenkawa> I recommend this cooler if anyone needs one (waveshare's VF2 cooler)
<mps> Tenkawa: with your cmdline params I've got "bootarg overflow 284+0+0+1 > 256" from u-boot
<mps> have to look at u-boot where to increase this
<Tenkawa> then you need to go increase your u-boot max string length
<Tenkawa> yeah
<mps> Tenkawa: didn't help with your cmdline params
<Tenkawa> mps: not sure what to tell you then... the only variable really is mainline u-boot from anyone else so far I've seen doing testing here.
<mps> Tenkawa: did you tested it with mainline u-boot
<Tenkawa> No
<Tenkawa> I don't have the source url.. I can try building another drive with it.. problem is soon as I use that u-boot it can't be used on this image without risking this staging so I'll need to replicate the drive
MaxGanzII_ has joined #riscv
MaxGanzII has quit [Remote host closed the connection]
Stat_headcrabed has quit [Quit: Stat_headcrabed]
<Tenkawa> Well things were going good until apt just scrambled my initrams
<Tenkawa> I won't be able to fi those until I get home
<Tenkawa> er fix
BootLayer has quit [Quit: Leaving]
valdemaras has joined #riscv
Nixkernal has joined #riscv
davidlt has quit [Ping timeout: 244 seconds]
Andre_Z has quit [Ping timeout: 246 seconds]
Tenkawa has quit [Quit: Was I really ever here?]
<mps> Esmil: with your last patches kernel boots very fine from nvme and mainline u-boot - U-Boot 2023.10-rc3-dirty (Aug 26 2023 - 17:40:26 +0000)
<Esmil> mps: cool. for some reason i still get the 'nvme0: I/O 640 QID 5 timeout, completion polled' errors. even with their latest u-boot release and the 6A powersupply that came with my vf2
<mps> Esmil: latest fixes for u-boot are few days old and there is one patch not yet applied
Kedleston has quit [Ping timeout: 255 seconds]
Andre_Z has joined #riscv
Kedleston has joined #riscv
Andre_Z has quit [Ping timeout: 245 seconds]
Andre_Z has joined #riscv
heat has joined #riscv
prabhakarlad has quit [Ping timeout: 246 seconds]
Andre_Z has quit [Quit: Leaving.]
ellis has joined #riscv
valdemaras has quit [Quit: valdemaras]
freakazoid332 has quit [Read error: Connection reset by peer]
freakazoid332 has joined #riscv
aerkiaga has quit [Remote host closed the connection]
jmdaemon has joined #riscv
hrberg has quit [Ping timeout: 256 seconds]
crabbedhaloablut has quit []
ntwk has quit [Quit: ntwk]
stolen has quit [Quit: Connection closed for inactivity]
junaid_ has quit [Remote host closed the connection]
ntwk has joined #riscv
junaid_ has joined #riscv
mwette has quit [Quit: ERC 5.4 (IRC client for GNU Emacs 28.2)]
ellis has quit [Remote host closed the connection]
junaid_ has quit [Remote host closed the connection]
Tenkawa has joined #riscv