2020-09-28

<montjoie> it issue a warning... I will fix thanks
<montjoie> megi: I check, I add a manual error and see what happen

2020-09-25

<montjoie> remains to be done, RSA need to mainlined
<montjoie> it remains RSA...
<montjoie> \o/
<montjoie> sun8i-ss/sun8i-ce RNG hash merged!

2020-09-20

<montjoie> oh no EMCE

2020-09-11

<montjoie> and ciphers does not work in BE...
<willmore> That was in sympathy for montjoie.
<montjoie> another hw bug bite the dust
<montjoie> ah ah seems A10/A13 has IV register fail to work
<montjoie> wens: just got a new comment, need v7...
<montjoie> so I wait next batch of apply from crypto maintainer
<montjoie> wens: not yet, I sent v6, and the only comment was on the last style patch
<wens> montjoie: has your sun8i-ce/ss rng series been accepted yet?

2020-09-10

<montjoie> well, broken and bypassed
<montjoie> OMG sun4i-ss is broken... since 04/19
<montjoie> Anyway even with hundred of boards power is already to be a thing to manage
<montjoie> gigOwatt exactly in my memory
<ElBarto> montjoie is french so that's 2.21 gigawatt for us :)
<montjoie> when I will have 100 000 board in the lab... I will need bigger power supply
<montjoie> boots not bot
<JohnDoe_71Rus> https://i.imgur.com/0OcUKS1.jpg montjoie and his bots
<montjoie> hé hé 100 000 sunxi boots in my kernelci lab \o/

2020-09-04

<mps> montjoie: yes
<montjoie> using sunxi_defconfig ?

2020-08-28

<montjoie> willmore: for reverse polarity story, I created a beautiful adapter, well solder (for me very rare), well protected and.... the machine which need it was the only DC barrel I know with + oustide
<willmore> montjoie, I'm looking at the M1+ schematic and it looks like it's just fine to power it from the GPIO. As usual, it has no reverse polarity protection there, so be careful.
<martinayotte> montjoie: Don't get fear, simply download schematic :-)
<montjoie> I am stingy and dont want to buy new devices:)
<montjoie> I fear to destroy it
<montjoie> anyway the urgent "give cartoons to child for still working" was fixed with a potato, solder dc barrel will wait
<montjoie> according to the datasheet, all VDD in the header are output only
<martinayotte> montjoie:my NanoPiM1+2 is powered via 5V GPIO header since months ...
<montjoie> I dont have a opipc+ only opipc, I will check this morning
<wens> montjoie: could you check the ethernet leds on both the opipc and opipc+ ?
<montjoie> does someone know if I can power a nanopim1+ via headers ? the microusb power is half unsolder
<montjoie> wens: I got one

2020-08-25

<montjoie> how many ?

2020-08-22

<montjoie> in fact only H6 has TRNG working
<montjoie> tuxd3v: SS has only PRNG
<montjoie> tuxd3v: yes latest is v5, for SS it is patch 1 to 6
<tuxd3v> montjoie, thanks
<montjoie> tuxd3v: and it should be applied this time
<wens> tuxd3v: there's a renewed series for RNG / hash support for the crypto engine by montjoie . see LAKML

2020-08-20

<montjoie> so it cannot be a lab issue
<montjoie> oh all 5.9-rc1 failling on pine64-plus
<montjoie> MoeIcenowy: let me retry the job on it
<montjoie> I confirm lof of panic on kernelci...

2020-08-14

<montjoie> ndufresne: a 3 year old patch, why didnt they upstream it!
<ndufresne> montjoie: diego71: thanks for the support, I managed to make it work by applying armbian patch, https://github.com/armbian/build/blob/master/patch/u-boot/u-boot-sunxi/board_lime2/GMAC_TX_DELAY_autocorrection.patch
<ndufresne> montjoie: yes, apparently, and clearly they got something that made it work
<montjoie> ndufresne: so the problem is uboot/net ?
<ndufresne> montjoie: diego71 thanks for the feeback, so I'll ask on #olimex, and try 2019.07 (was using latest 2020
<montjoie> ndufresne: and some fast check on kci I saw 5.8-rc1 working
<montjoie> ndufresne: at least uboot/ethernet work on it 2019.07 and linux 5.2.17

2020-08-07

<tuxd3v> montjoie, I mean more complete
<montjoie> tuxd3v: good shape ?

2020-07-25

<Ashleee> montjoie, what about that museum? :)
<mru> montjoie: this museum: https://www.museumtorture.com/
<montjoie> rz fixing 3.4 is a loss of time, it is too old. It belongs to a museum!

2020-07-22

<montjoie> I try to made it oworking
<montjoie> me too
<montjoie> jernej: my problem seems different, enabling EMCE prevent emmc(mmc2) to work
<jernej> montjoie: As I suggested to Ashleee already yesterday, try adding "mmc-hs200-1_8v;" or "mmc-hs400-1_8v;" property to emmc node.
<montjoie> probably it is the EMCE which do that....
<montjoie> mmmh I check all other jobs which ran on it, emmc always here
<montjoie> I mean it is the kernel which dont detect it
<montjoie> anyway it is cold boot always
<montjoie> okay just break the pattern with current boot..
<montjoie> not enough boot to say, but the pattend seems work notwork work notwork etc
<montjoie> what can I do ?
<montjoie> hello on my sun50i-h6-pine-h64-model-b, the emmc is detected only 1 boot on 2

2020-07-14

<montjoie> in my testsuite I play modprobe all then rmmod all. lot of fun
<montjoie> yes, lot of drm module are not rmmod-proof

2020-07-12

<montjoie> willmore: and just after building everything, I saw better for same price...
<willmore> montjoie, congrats.
<montjoie> ah ah a new ryzen+nvme, compile testing the kernel will be so fast

2020-07-10

<montjoie> it works perfectly non-H6+BE
<montjoie> but this case seems insane
<montjoie> I love to have a portable driver
<MoeIcenowy> montjoie: why is there BE?
<lvrp16> montjoie: where head meets the desk
<montjoie> but this conversion is needed only for RNG and hashes, not ciphers
<montjoie> perhaps the start for calling cthlhu
<montjoie> this is INSANE
<montjoie> H6 CryptoEngine convert BE to LE but only for some part of the DMA descriptor
<montjoie> when you rediscover bugs you found 3 month ago and forgoted them...

2020-07-09

<montjoie> datasheets ?

2020-07-08

<diizzy> montjoie: cp2xxx as suggested earlier as a better choice, you can find "cheap" ftdi stuff even in eu that's geniune
<montjoie> diizzy: yes I know what I can get from aliexpress, my worst xp was a counterfeit of a counterfeit USB net dongle
<diizzy> montjoie: ftdi stuff on aliexpress is probably not genuine
<montjoie> but I found some which have it
<montjoie> yes cheapest one does not have the serialID, which is a requirement
<montjoie> but I found FTDI on aliexpress without thecable (and cheaper)
<montjoie> anyway, FTDI are 15€, cuting them is hard
<montjoie> you mean ®™© ? ALTGR+shift+c/r/8
<montjoie> KotCzarny: maked cable are potential less reliable
<montjoie> note that it is planned to create a hole between two stack and hide cable below the wood
<fALSO> LOL montjoie
<montjoie> I want shorter cable!!
<montjoie> found an old photo kernel.montjoie.ovh/lab-3.jpg
<montjoie> gediz539: the work one is better (and more large), I need to pick a photo of it
<gediz539> Ashleee, montjoie: cool labs :)
<montjoie> I still didnt took the time to properly set boards in the right place
<montjoie> note that I need to publish the updated lab 4
<montjoie> KotCzarny: hé hé hé

2020-07-03

<montjoie> diizzy: if you dont have comment here, try the mailing list, you can send the patch as RFC/RFT
<montjoie> diizzy: this is the right place

2020-06-29

<gediz0x539> montjoie: 'll give it a try, thanks.
<montjoie> gediz0x539: mutt

2020-06-26

<montjoie> nobody will know it is
<montjoie> paste it
<montjoie> you could sticker it
<montjoie> this kind of password has many advantages, it is good AND it do not look like a password
<montjoie> I repeat my password hints, use a chained cmdline, like smarctl -a /dev/sda && touch /tmp/test
<montjoie> wens: not bad but could be better

2020-06-21

<montjoie> for opi3 I need to made it work on linux first
<montjoie> jernej: for 1+ I tried to add MACPWR without result
<montjoie> in your branch
<montjoie> jernej: but I dont see anything for one+/opi3 net
<montjoie> I use my pineh64 with it
<montjoie> jernej: you coul rebase on master, "net: sun8i_emac: Support for H6 SoC" is merged
<jernej> montjoie: I'm using H6 eth U-Boot patch from smaeul and my own hack so I can netboot OPi3
<montjoie> is seems opi3 does not have ethernet in linux..
<montjoie> oups I forgot to add "in uboot"
<montjoie> does someone made eterhnet works on opi3/opi1+ ?

2020-06-13

<montjoie> ok found it
<montjoie> mru: how to use ?

2020-06-12

<montjoie> all returning 0xff
<montjoie> I have a good target sun4i-a10-olinuxino-lime with 50: 50 51 52 53 54 55 56 57 -- -- -- -- -- -- -- --
<montjoie> mru: if you know the exact version of busybox I can use it
<montjoie> I have send jobs with arm64 only, will try arm
<montjoie> on what board do you trigger it ?
<montjoie> but yes I use official i2c-tools
<montjoie> on a meson-sm1-sei610
<montjoie> Dump I2C bus 0 address 18 => 0x50
<montjoie> I have scripted a bit more my i2c test, I will brodcast them https://github.com/montjoie/lava-tests/blob/master/i2c.sh
<montjoie> done on a pineh64
<montjoie> 0x08
<montjoie> i2cdump -f -y 0 0x51 i works for example
<montjoie> with a device no crash
<montjoie> "Error: Block read failed, return code -6" no crash
<montjoie> anyway I will see it soon
<montjoie> mru: does address 51 works also ?
<montjoie> I need time to refind which one has it
<montjoie> what kernel version do you need ? next is ok ?
<montjoie> I have sensors or eeprom does it change something ?
<montjoie> so what bug ?
<montjoie> I have some (and anyway I need to buy cheap i2c anything to wire on all boards)
<montjoie> I have lot of boards to bug
<montjoie> lets try

2020-06-08

<montjoie> wow I just saw that orangepi-pc was added to qemu

2020-05-30

<montjoie> 3gb max in my memory

2020-05-28

<montjoie> speaked too fast it works
<montjoie> anyway the new uboot dont work
<montjoie> some old .pyc (for python2) file remaining
<montjoie> ok found the issue
<montjoie> it is the less painfull of remaining board to add
<montjoie> I have a C.H.I.P and try to add it to kernelci
<montjoie> binman: bad magic number in 'binman.etype': b'\x03\xf3\r\n'
<montjoie> anyone recently compiled uboot for the chip ? it fail

2020-05-25

<montjoie> ah ah an user of mc! or just a bad pass...
<montjoie> it belongs to a museum!

2020-05-19

<montjoie> but he was so smart that the society shut down just after my leave
<montjoie> exactly my boss didnt give me the choice
<montjoie> exa
<willmore> montjoie, we've all done embarrasing things for money. :)
<montjoie> ah ah hello word die due no ENOMEM
<montjoie> yeah, my school wanted we do java on 486...
<montjoie> I teached word and excel also, yeah young and ENOMONEY
<montjoie> server2003 on celeron... e^hurts
<montjoie> I teached people to install server2003, it still hurts
<montjoie> you use windows without even recompiling its kernel for being more rude
<montjoie> KotCzarny: just say I use windows, this is the ultimate insult

2020-05-15

<montjoie> the OS was some realtime, didnt remember the name
<montjoie> running linux on a 4x486
<montjoie> it was my dream
<montjoie> ultimate class!
<montjoie> my first stage was in a ciment industry, everything was controled by some MULTI 486 MB
<willmore> montjoie, No, I don't think I should. I have a feeling I'll lose the rest of the day to it. ;)
<montjoie> hardcore, try v4v
<montjoie> at least I started with a VGA
<willmore> montjoie, too bad I only used herculese mono at that time--too poor for a color monitor. Heck, my next monitor was an NEC 2GS mono VGA monitor. At least I could play graphical games--even though I got not color.
<montjoie> omg they speak!
<montjoie> willmore: I remember playing wolf3d with PC speaker, and suddently THE GIFT, a soundblaster
<willmore> montjoie, yes it was!
<montjoie> never forget IO, cores without high speed IO disk is useless
<montjoie> but it is enough for doom and wolf3d
<montjoie> YEAH

2020-05-14

<montjoie> damex: try first boot tftp with a ramdisk, then when it works try nfs

2020-05-11

<montjoie> and perhaps in the future it will have the embedded crypto for emmc working
<montjoie> pineh64 at least

2020-05-10

<montjoie> related to the budget you can invest in it
<montjoie> the hard part is reliable PDU
<montjoie> for each board you need serial and a PDU
<damex> montjoie: thanks, will read that. i didn't knew there were already something for that purpose :)
<montjoie> damex: you could read kernelci documentation
<montjoie> damex: for testing builds on real sbcs, you should use LAVA
<montjoie> systemd with 3.4 is one of the way to invoke satan

2020-05-08

<sunshavi> montjoie: I have found a bug on the one that is mainlined. I am with kernel 5.6.7
<montjoie> sunshavi: the "new" driver has bug and should not be used, the correct one is in mainline
<sunshavi> montjoie: how could I try the new dwmac driver?
<montjoie> ah the joy to create a totaly new driver and a week before merging, finding it must be abandoned and merged in dwmac-sun8i

2020-05-07

<willmore> montjoie, was that you who rant into the obscured ethernet?

2020-05-03

<montjoie> perhaps armbian ?
<montjoie> cardboard64: https://linux-sunxi.org/Linux_mainlining_effort seems supported
<montjoie> cardboard64: see the support matrix

2020-04-28

<sunshavi> montjoie: https://termbin.com/wloh
<montjoie> do you have any kernel/log message?
<montjoie> does the status goes to down ?
<sunshavi> montjoie: ^^^
<montjoie> sunshavi: define more precisely "drops"

2020-04-21

<montjoie> you need to find the uboot.scr/uboot.cmd
<montjoie> probably in /boot/
<montjoie> you dont need, you need to copy the dtb
<montjoie> many dtb should be in that directory, but you need sun50i-h6-pine-h64-model-b.dtb
<montjoie> in arch/arm64/boot/dts/allwinner/
<montjoie> you should get a dtb file after kernel compile
<montjoie> their build system should know that a new dt exists for 5.6
<montjoie> I dont use armbian so dont know, I think a bug should be open at armbian
<sbbg> montjoie, but that's weird. I assign pineh64_b when I run compile.sh. Is there other way to impose the DT for Armbian build system?
<montjoie> sbbg: yes you need the sun50i-h6-pine-h64-model-b
<sbbg> montjoie, to my surprise... dmesg|grep -i model tells "Machine model: Pine H64 Model A"

2020-04-20

<kilobyte_ch> montjoie: I need tvin for my application and havent' yet ported it to something newer
<montjoie> 3.4.75 is really old (and insecure)
<montjoie> kilobyte_ch: why not trying more recent kernel ?
<montjoie> sbbg: dmesg |grep -i model
<montjoie> sbbg: with dmesg what machine model it give ?
<montjoie> sbbg: which DTB do you use ?
<sbbg> montjoie: Hi, I did try with PineH64 model B with latest Armbian kernel 5.6.0, both ethernet and Wifi are down. While older 5.5.6 has wifi working.

2020-04-18

<montjoie> for A80, only few devices exists like cubie4

2020-04-17

<montjoie> anyway the modelB is in 5.6... to late to chaneg anything
<montjoie> DT
<montjoie> jernej: sbbg: check for the new model B
<montjoie> oh yes I forgot the introducing of model B
<MoeIcenowy> montjoie: I think an origin of issue may be that sun50i-h6-pine-h64 is originally for model B
<montjoie> perhaps I need to test model B
<montjoie> interesting I tested mode A yesterday
<sbbg> montjoie, thank you. It's model-B for me.
<montjoie> it is two people complaining about 5.6 h6 and ethernet, I will check
<montjoie> my last renesas uboot update is just some week ago, still have injury
<montjoie> WORSE exactly
<montjoie> at least for sunxi, updating uboot is easy, just try some renesas to know what the hell is
<montjoie> for sunxi, uboot network is reliable, this is good
<montjoie> and yes you need dhcp + tftp
<montjoie> damex: yes you need uboot to be on sdcard
<damex> montjoie: well, i would prefer built-in ethernet. so netbooting with h5 like described above - still needs uboot to be loaded from local media?
<montjoie> damex: for H- you still can use a net usb dongle
<montjoie> damex: fo h6 only pineh64 work (but you need to patch uboot)
<montjoie> damex: h5 is fully usable via netboot

2020-04-16

<montjoie> jernej: nothing on my side for dwmac-sun8i
<ElectroStrong> thx montjoie - I'll check it out
<montjoie> ElectroStrong: I see sun4i-a10-hackberry in mainline linux, so any version should be fine
<montjoie> jernej: 5.6.0 or any 5.6.x ?
<montjoie> jernej: let me check
<jernej> montjoie: Did you noticed any issue with dwmac-sun8i on 5.6.0 (H6)?

2020-04-13

<montjoie> tuxd3v since i know your are interest in, I just sent TRNG patches

2020-04-10

<megi> montjoie: enabling the ethernet power properly, I'd guess
<montjoie> smaeul: for H6 uboot network, what prevent opi3+ to work?

2020-04-09

<montjoie> sbbg: try stracing to see what it do
<montjoie> psydread: try reading the brief of the mmc controller
<montjoie> never heard of a limit, but this is not sufficiant
<montjoie> if you fear to have error message try dmesg -l err dmesg -l warn dmesg -l crit
<montjoie> sbbg: i dont see any errror message
<montjoie> wow SS hashes are broken on a80

2020-04-08

<montjoie> ouch error festival on A80
<montjoie> youhou, I finally booted in kernelci this ... of cubieboard4 (via a rootfs which fake a uboot and then kexec the final kernel)

2020-04-05

<montjoie> thanks, I will save lots of USB dongle
<montjoie> smaeul: your H6 network patch for uboot WORKS!!

2020-04-04

<montjoie> having the same kernel for too long is a bad sign of security update
<Kasreyn> montjoie: OK.. thanks :)
<montjoie> Kasreyn: the empty IV problem was fixed later
<montjoie> Kasreyn: such old kernel, update it

2020-04-03

<montjoie> and I am polishing the patch serie
<montjoie> tuxd3v: it exists
<wens> tuxd3v: TRNG IIRC is not supported yet, but you can ask montjoie

2020-04-01

<montjoie> hello does "sun4i-usb-phy 1c19400.phy: Couldn't request ID GPIO" is serious or not ? on A33

2020-03-27

<montjoie> smaeul: where is this fix for opi3 ?
<montjoie> tuxd3v: A20 SS has PRNG
<smaeul> montjoie: the ATF changes for H6 have all been merged to mainline (except for fixing Orange Pi 3)

2020-03-26

<montjoie> smaeul: nothing to add in ATF ?

2020-03-25

<montjoie> does someone has a link to latest uboot H6 network patchs?

2020-03-21

<montjoie> ah ah you copy all firmware, but yours has 8G RAM, all firmware are different
<montjoie> move switch #1, move switch #2 , move switch #3, then copy paste something which is wrong
<montjoie> renesas I see you, 1 hour to upgrade firmware
<montjoie> ndufresne: clearly, I have installed a lot of board in our lab, and sunxi is at top...
<willmore> montjoie, good call on the duct tape
<montjoie> if the crypto driver said "failed to fejfjffefjhjkh" one day, you will know the reason
<montjoie> my little son is inserting random caracther when I code, it is like chaos coding
<tuxd3v> montjoie, :)
<montjoie> tuxd3v: with 3 child, 1 week is already long, but instead of buying pasta and toilet paper, I have bough all duck tape for neutralize them ... priority is important
<tuxd3v> montjoie, partial confinment is nice, for a long long time, it would be painfull..
<montjoie> I love confinment, so many time to work on patchs