ChanServ changed the topic of #linux-rockchip to: Rockchip development discussion | IRC log http://irclog.whitequark.org/linux-rockchip | Community GH https://github.com/linux-rockchip | Rockchip GH https://github.com/rockchip-linux | ML https://groups.google.com/group/linux-rockchip
tllim has quit [Quit: Leaving]
muvlon_ has quit [Ping timeout: 265 seconds]
nighty- has joined #linux-rockchip
muvlon_ has joined #linux-rockchip
anarsoul|3 has quit [Ping timeout: 240 seconds]
cnxsoft has joined #linux-rockchip
yann|work has quit [Ping timeout: 256 seconds]
yann|work has joined #linux-rockchip
vstehle has quit [Ping timeout: 244 seconds]
dejavou42 has joined #linux-rockchip
lurchi_ is now known as lurchi__
dejavou42 has left #linux-rockchip [#linux-rockchip]
DunningFreddyKru has quit [Quit: Page closed]
lurchi_ has joined #linux-rockchip
lurchi__ has quit [Ping timeout: 240 seconds]
vstehle has joined #linux-rockchip
amstan has quit [Ping timeout: 276 seconds]
vstehle has quit [*.net *.split]
kaspter has quit [*.net *.split]
_whitelogger has joined #linux-rockchip
daniels has joined #linux-rockchip
seanpaul has joined #linux-rockchip
norris has joined #linux-rockchip
ojn has joined #linux-rockchip
lvrp16 has joined #linux-rockchip
steev has joined #linux-rockchip
leah2 has joined #linux-rockchip
ayaka has quit [Ping timeout: 260 seconds]
ayaka has joined #linux-rockchip
<Ke> what are the best speeds that people have gotten with uSD on rk3288-veyron-speedy and rk3399-kevin
<Ke> I have super fast samsung uSD cards and have never reached anything above 30MB/s
<Ke> nominal read write speeds should be around 100MB/s
xerpi has joined #linux-rockchip
yann|work has quit [Ping timeout: 260 seconds]
matthias_bgg has joined #linux-rockchip
sb35 has quit [Ping timeout: 268 seconds]
athidhep has joined #linux-rockchip
athidhep has quit [Changing host]
athidhep has joined #linux-rockchip
sb35 has joined #linux-rockchip
athidhep has quit [Ping timeout: 240 seconds]
perr has joined #linux-rockchip
perr has joined #linux-rockchip
yann|work has joined #linux-rockchip
<fysa> SD clock is limiting them I believe
<fysa> quoted speeds on the cards are for 100mhz clocks I think?
<fysa> and I think ours are running quite a bit lower
<fysa> you might be able to overclock with a dt overlay or something
<phh> I think you should have 100Mhz 4bit bus, so ~ 50MB/s?
<phh> I don't think mmc protocol has such a big overhead to fall to 30MB/s
<phh> and afaik mmc clocks are power of 2 multilpes of 25MHz (or 12.5?)
perr has quit [Read error: Connection reset by peer]
<fysa> I see around 100MB/s on eMMC and 30MB/s on uSD on rk3328
<fysa> ah, SD card on rock64 is limited to DDR50 mode
<fysa> not sure about others
<Ke> fysa: so there is no hw limitation in getting faster bus speeds?
lukasz_ has joined #linux-rockchip
<hgs> how to extract contents of rk3399 boot.img?
<wens> seems rk3288 and rk3399 dont have UHS-1 enabled for uSD
<wens> highspeed is only 25 MB/s
<Ke> wens: could just patching the device tree help then?
<Ke> if so, why is it not on by default
xerpi has quit [Quit: Leaving]
<Ke> I guess I could take a look at the dwmmc
<Ke> I'm a big boy and know how to read kernel
<wens> I don't know, I'm not familiar with rockchip stuff
<eballetbo> Ke: device tree (at least for kevin) has the following modes enabled
<eballetbo> sd-uhs-sdr12;
<eballetbo> sd-uhs-sdr104
<eballetbo> sd-uhs-sdr50;
<eballetbo> sd-uhs-sdr25;
indy has quit [Quit: ZNC - http://znc.sourceforge.net]
<eballetbo> i don't think is a dt problem, what's the speed detectec by the kernel for your sd-card?
<wens> the driver code does not has HS or UHS enabled by default. The device tree properties do that.
<wens> if the modes aren't enabled, the kernel won't even try them
<wens> looks like they are enabled for some boards, like kevin, veyron, but not others
<eballetbo> ooi, how do you exactly test the speed? just to run the same test here
<Ke> eballetbo: standard dd if=... of=/dev/mmcblk0p2 bs=1M conv=fsync
<Ke> cached writes with flush
<Ke> eballetbo: if you can get better speed at any test at all, please let me know
<Ke> any technically correct test above 40MB/s is definitely of interest
<Ke> or even 30, I am not quite sure I have not exaggerated a bit
<Ke> I think typical is about 27
<eballetbo> Ke: sure, I can test this afternoon
<Ke> thanks
ramcq has quit [Quit: brb]
<eballetbo> not sure about the results :P ,but I have interest on it
<Ke> eballetbo: this is the dmesg from veyron-speedy, I keep seeing 148500000Hz and SDR104 a lot
<Ke> I am not sure how to figure out the speeds as such
<Ke> IIRC it was similar on Kevin
<Ke> have to check later today
<wens> Ke: SDR104 means top speed is 104 MB/s
ramcq has joined #linux-rockchip
<eballetbo> Ke: My sdcard is detected as
<eballetbo> [ 2.922826] mmc0: new ultra high speed SDR50 SDHC card at address 0001
<eballetbo> and running the following I get
<eballetbo> root@debian:~# dd if=/dev/urandom of=rnd_data bs=1024 count=1024k conv=sync
<eballetbo> 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 39.2514 s, 27.4 MB/s
<eballetbo> 1048576+0 records out
<eballetbo> 1048576+0 records in
<Ke> seems to be exactly in line with the "(13:19:15) < Ke> I think typical is about 27"
<eballetbo> I suppose that if your card is detected as SDR104 then should be faster.
<Ke> we have measure about 60MB/s at work with similar cards
<Ke> I think also my usb-sd adaptor does
<Ke> having string SDR104 does not a priori imply the bus speed is SDR104, especially, when the clock speed is 148500000Hz
<Ke> https://users.aalto.fi/~jkarlson/files/dmesg.txt apparently this log got omitted
<Ke> "UHS104 (SDR104) also support a clock frequency of 208 MHz, which could transfer 104 MB/s."
<Ke> even if the bus speed is high, the host controller could be still limited to lower speeds ar DMA side or something
<wens> debugfs should report the actual bus speed in /sys/kernel/debug/mmcN/ios
<eballetbo> yeah, also dmesg shows it, my bus speed is 100MB (matches with a SDR50 type card), so if your SD-card is faster and the clock speed is faster you should get faster results I guess. We're missing something
<eballetbo> not sure if I have a UHS104 card here
<Ke> wens: thanks
<Ke> does mmc2 correspond to mmcblk2?
indy has joined #linux-rockchip
<Ke> I guess readlink /sys/block/mmcblk0/device
<Ke> anyway it's mmc0 and timing spec:6 (sd uhs SDR104)
nighty- has quit [Quit: Disappears in a puff of smoke]
<Ke> what about that dma thing? is it expected that dma by a controller is way faster?
xypron has quit [Remote host closed the connection]
sb35 has quit [Ping timeout: 276 seconds]
<wadim_> Ke: at what freq. is your controller running?
<Ke> clock:150000000 Hz
<Ke> the same details as I get in dmesg
<wadim_> on a rk3288?
<Ke> yes, I don't have Kevin at hand atm
<wadim_> I never saw a rk3288 device running with up to 150Mhz :/ interesting it is working for you
<Ke> if by working you mean, I get text on debugfs and kernel log, sure
<Ke> it's still slow
<wens> what about read performance?
cnxsoft has quit [Read error: Connection reset by peer]
cnxsoft has joined #linux-rockchip
<hgs> hgs: hi all, need help adapting uboot defconfig & creation of boot.img to make rk3399 boot to a linux distro (buildroot) (https://github.com/rockchip-linux)
<Ke> hmm, somethin weird, I get 63.5 MB/s
<Ke> dd iflag=direct bs=100M count=1 if=/dev/mmcblk0 of=/dev/null
<Ke> also with 200M
<eballetbo> more than me ...
<eballetbo> # dd iflag=direct bs=500M count=1 if=/dev/mmcblk0 of=/dev/null
<eballetbo> 524288000 bytes (524 MB, 500 MiB) copied, 11.0619 s, 47.4 MB/s
<eballetbo> 1+0 records out
<eballetbo> 1+0 records in
<eballetbo> It makes sense as you sd-card and bus speed is faster
<Ke> hmm apparently my writes were too small originally
<Ke> also I had sudo blkdiscard "$disk"
<Ke> without flush
<Ke> disk=/dev/disk/by-partlabel/SD_kernel; sudo blkdiscard "$disk" && sync && sudo dd if=/dev/zero "of=${disk}" bs=32M conv=fsync oflag=direct
<Ke> 33537536 bytes (34 MB, 32 MiB) copied, 0.596419 s, 56.2 MB/s
xypron has joined #linux-rockchip
<Ke> this is the most common benchmark I did when updating the kernel, not very good benchmark
<Ke> still this is moderately slow, but a lot better than before
<eballetbo> I don't think this is slow, assuming your clk is at 150MHz you can get maximum 75MHz, 56.2 MB/s doesn't seem too bad for me. Well, these sd test always confuses me with the results :P
<eballetbo> s/75MHz/75MB/
<Ke> ok
<Ke> thanks everyone
<Ke> I am almost sure I have benchmarked this properly at some point
<Kelsar> hmm... how do i get my system image onto the nvme connected to the arm board...
XingZheng has joined #linux-rockchip
<hgs> hi all, should i be concerned about the follow message "Trasfer Length NOT set, please use new version tool" in the console while flashing the images
cnxsoft has quit [Read error: Connection reset by peer]
cnxsoft has joined #linux-rockchip
afaerber has quit [Quit: Leaving]
afaerber has joined #linux-rockchip
athidhep has joined #linux-rockchip
athidhep has joined #linux-rockchip
athidhep has quit [Changing host]
tllim has joined #linux-rockchip
lurchi_ is now known as lurchi__
yann|work has quit [Ping timeout: 240 seconds]
amstan_ has quit [Quit: ZNC - http://znc.in]
amstan has joined #linux-rockchip
amstan has joined #linux-rockchip
amstan has quit [Changing host]
tllim has quit [Quit: Leaving]
gsh has joined #linux-rockchip
yann has joined #linux-rockchip
cnxsoft has quit [Quit: cnxsoft]
JohnDoe_71Rus has joined #linux-rockchip
Kelsar has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
Kelsar has joined #linux-rockchip
yann has quit [Ping timeout: 244 seconds]
yann has joined #linux-rockchip
ayaka has quit [Ping timeout: 240 seconds]
anarsoul|3 has joined #linux-rockchip
anarsoul has quit [Ping timeout: 248 seconds]
ayaka has joined #linux-rockchip
anarsoul|3 has quit [Ping timeout: 240 seconds]
JohnDoe7 has joined #linux-rockchip
matthias_bgg has quit [Quit: Leaving]
JohnDoe_71Rus has quit [Ping timeout: 240 seconds]
gsh has quit [Ping timeout: 252 seconds]
hanetzer has quit [Ping timeout: 244 seconds]
hanetzer has joined #linux-rockchip
anarsoul|3 has joined #linux-rockchip
yann has quit [Ping timeout: 255 seconds]
vagrantc has joined #linux-rockchip
anarsoul|2 has joined #linux-rockchip
anarsoul|3 has quit [Ping timeout: 260 seconds]
kaspter has quit [Read error: Connection reset by peer]
kaspter has joined #linux-rockchip
anarsoul|2 has quit [Ping timeout: 240 seconds]
anarsoul|2 has joined #linux-rockchip
unics has joined #linux-rockchip
unics has quit [Quit: Leaving]
Easyfab has joined #linux-rockchip
JohnDoe7 has quit [Quit: KVIrc 5.0.0 Aria http://www.kvirc.net/]
tllim has joined #linux-rockchip
Easyfab has quit [Quit: Leaving]
lurchi__ is now known as lurchi_
lukasz_ has quit [Ping timeout: 240 seconds]
lukasz has joined #linux-rockchip
LargePrime has quit [Remote host closed the connection]
LargePrime has joined #linux-rockchip
Substring has joined #linux-rockchip
lurchi_ is now known as lurchi__
lurchi__ is now known as lurchi_
afaerber has quit [Quit: Leaving]
afaerber has joined #linux-rockchip
lukasz has quit [Quit: Leaving]
Substring has quit [Quit: Leaving]
anarsoul|3 has joined #linux-rockchip
anarsoul|2 has quit [Ping timeout: 255 seconds]
vstehle has quit [Ping timeout: 240 seconds]