Whats that change doing? enabling or disabling 64bit inodes? Did the arm64 libraries for mkfs.xx (and friends, fsck etc…) finally catch up to amd64 or something?
the exact content of the change can be displayed on github, there were 3 commits in the past days;
I have rebuilt my firmware, by using
building the img and flashing the EMMC seemed to be as successfull as before
but , again, I got the following errors (in bold case) at boot time:
BULKcmd[ burn_complete 1] reboot mode: normal bl31 reboot reason: 0xd bl31 reboot reason: 0x1 system cmd 1. GXL:BL1:9ac50e:a1974b;FEAT:ADFC318C;POC:3;RCY:0;EMMC:0;READ:0;0.0;CHK:0; TE: 127612 BL2 Built : 20:32:17, Sep 8 2017. gxl g6296b83 - xiaobo.gu@droid12 set vcck to 1120 mv set vddee to 1000 mv Board ID = 6 CPU clk: 1200MHz DQS-corr enabled DDR scramble enabled DDR3 chl: Rank0+1 @ 768MHz - PASS Rank0: 1024MB(auto)-2T-11 Rank1: 1024MB(auto)-2T-11 DataBus test pass! AddrBus test pass! -s Load fip header from eMMC, src: 0x0000c200, des: 0x01400000, size: 0x00004000 New fip structure! Load bl30 from eMMC, src: 0x00010200, des: 0x01100000, size: 0x0000d600 Load bl31 from eMMC, src: 0x00020200, des: 0x05100000, size: 0x0002c600 Load bl33 from eMMC, src: 0x00050200, des: 0x01000000, size: 0x00066a00 NOTICE: BL3-1: v1.0(release):a625749 NOTICE: BL3-1: Built : 11:25:15, Aug 25 2017 [BL31]: GXL CPU setup! NOTICE: BL31: BL33 decompress pass mpu_config_enable:ok [Image: gxl_v1.1.3243-377db0f 2017-09-07 11:28:58 qiufang.dai@droid07] OPS=0x82 9d 42 bd c7 a1 12 ec 2f 5f 6 30 b8 [0.354484 Inits done] secure task start! high task start! low task start! ERROR: Error initializing runtime service opteed_fast U-Boot 2015.01 (Jun 04 2018 - 21:29:41) DRAM: 2 GiB Relocation Offset is: 76eae000 register usb cfg = 0000000077f56a18 [CANVAS]canvas init boot_device_flag : 1 Nand PHY Ver:1.01.001.0006 (c) 2013 Amlogic Inc. init bus_cycle=6, bus_timing=7, system=5.0ns reset failed get_chip_type and ret:fffffffe get_chip_type and ret:fffffffe chip detect failed and ret:fffffffe nandphy_init failed and ret=0xfffffff1 MMC: aml_priv->desc_buf = 0x0000000073eae6b0 aml_priv->desc_buf = 0x0000000073eb09d0 SDIO Port B: 0, SDIO Port C: 1 emmc/sd response timeout, cmd8, status=0x1ff2800 emmc/sd response timeout, cmd55, status=0x1ff2800 init_part() 293: PART_TYPE_AML [mmc_init] mmc init success dtb magic edfe0dd0 start dts,buffer=0000000073eb3200,dt_addr=0000000073eb3200 parts: 3 00: logo 0000000002000000 1 01: ramdisk 0000000002000000 1 02: rootfs ffffffffffffffff 4 init_part() 293: PART_TYPE_AML eMMC/TSD partition table have been checked OK! check pattern success mmc env offset: 0x6c00000 In: serial Out: serial Err: serial ## defenv_reserve Saving Environment to aml-storage... mmc env offset: 0x6c00000 Writing to MMC(1)... done reboot_mode=normal [store]To run cmd[emmc dtb_read 0x1000000 0x40000] _verify_dtb_checksum()-909: calc a2d5c9ca, store a2d5c9ca _verify_dtb_checksum()-909: calc a2d5c9ca, store a2d5c9ca dtb_read()-1031: total valid 2 dtb_read()-1098: do nothing vpu: clk_level in dts: 7 vpu: set clk: 666667000Hz, readback: 666660000Hz(0x300) vpu: vpu_clk_gate_init_off vpp: vpp_init hpd_state=0 cvbs performance type = 6, table = 0 Net: dwmac.c9410000amlkey_init() enter! [EFUSE_MSG]keynum is 4 [BL31]: tee size: 0 upgrade_step=2 [OSD]load fb addr from dts [OSD]fb_addr for logo: 0x7e000000 [OSD]load fb addr from dts [OSD]fb_addr for logo: 0x7e000000 [CANVAS]addr=0x7e000000 width=3840, height=2160 amlkey_init() enter! amlkey_init() 71: already init! [EFUSE_MSG]keynum is 4 [BL31]: tee size: 0 [KM]Error:f[key_manage_query_size]L507:key[usid] not programed yet [KM]Error:f[key_manage_query_size]L507:key[mac] not programed yet [KM]Error:f[key_manage_query_size]L507:key[deviceid] not programed yet saradc - saradc sub-system Usage: saradc saradc open <channel> - open a SARADC channel saradc close - close the SARADC saradc getval - get the value in current channel saradc get_in_range <min> <max> - return 0 if current value in the range of current channel gpio: pin GPIOAO_2 (gpio 102) value is 1 Product checking: pass! Hit Enter or space or Ctrl+C key to stop autoboot -- : 0 card out emmc/sd response timeout, cmd8, status=0x1ff2800 emmc/sd response timeout, cmd55, status=0x1ff2800 emmc/sd response timeout, cmd1, status=0x1ff2800
Bad device mmc 0 **
cfgload: fatload: no boot.ini or empty file
** File not found /boot/boot.ini **
cfgload: ext4load: no boot.ini or empty file
** File not found zImage **
** File not found uInitrd **
** File not found dtb.img **
** [rsvmem] get fdtaddr NULL!
** rsvmem - reserve memory**
Usage: rsvmem check - check reserved memory rsvmem dump - dump reserved memory rsvmem check failed Bad Linux ARM64 Image magic! InUsbBurn [MSG]sof Set Addr 23 Get DT cfg Get DT cfg set CFG
I suspected my context to be a bit unusual,
so I did
I was surprised gcc (and maybe other packages) were reinstalled ! this is unexpected and possibly wrong.
Let’s now wait a bit to finish the “make” step
same issue as before !
Failure here on buntu 17
This change disable ext4
metadata_csum feature, because current u-boot does not support such features.
The effect of not using 64bit ext4 is that you’ll be limited to ~ 15T volumes. Where as you can have 1EiB volumes if you use the 64Bit flag. However, 16T is the recommended max volume size for ext4 anyway.
Oh, it’s too bad…
I don’t have Ubuntu 17.10 build host.
It’s recommend to use Ubuntu 18.04. Ubuntu 17.10 is not a LTS and will end of life soon(July 2018).
You probably did not event try to take into account ubuntu 17 in your fixes,
so I did this change to your latest build-rootf file :
info_msg “Preparing image file for rootfs…”
if [ $(lsb_release -sc) == “bionic” ]; then
EXT4_MKOPTS="-q -F -L ROOTFS -m 2 -O ^64bit,^metadata_csum"
elif [ $(lsb_release -sc) == “artful” ]; then
EXT4_MKOPTS="-q -F -L ROOTFS -m 2 -O ^64bit,^metadata_csum"
elif [ $(lsb_release -sc) == “xenial” ]; then
EXT4_MKOPTS="-q -F -L ROOTFS -m 2"
then I removed all .img files from my previous make session
and then make
then when the resulting img was regenerated, if flashed as usual, and this time, it booted ok on vim1 pro !
and says this…
Debian GNU/Linux 9 Khadas ttyS0 Khadas login: khadas Password: The programs included with the Debian GNU/Linux system are free software; the exact distribution terms for each program are described in the individual files in /usr/share/doc/*/copyright. Debian GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent permitted by applicable law. khadas@Khadas:~$ uname -a Linux Khadas 4.9.40 #2 SMP PREEMPT Mon Jun 4 22:38:05 CEST 2018 aarch64 GNU/Linux khadas@Khadas:~$ [ 62.462624@3] fb: mem_free_work, free memory: addr:6af000 khadas@Khadas:~$
How about that ?
my change surely is much easier and more sustainable than your suggestion to change from buntu 17 to 18.
ubuntu 18.04 still need to be challenged and its all its undiscovered issues and quirks need some time before developpers massively jump into it
Now, let’s check which functions this uboot/kernel/linux combo are ok are which are not…
Never give up, never settle !
Ahh, yes, it’s my fault…forget to change it for Ubuntu 17.10. Sorry…
Yes, you are right, thanks for your work.
so here what my vim1 home built debian says…
khadas@Khadas:~$ uname -a
Linux Khadas 4.9.40 #2 SMP PREEMPT Mon Jun 4 22:38:05 CEST 2018 aarch64 GNU/Linu
khadas@Khadas:~$ lsb_release -a
No LSB modules are available.
Distributor ID: Debian
Description: Debian GNU/Linux 9.4 (stretch)
Great. I will fixup this or you can open a PR for this.
Welcome code donation
please just include those 2 lines, no need for a complicated PR, approval, merge, etc…
i’d like to use your script to build ubuntu 17 and 18 based firmwares,
currently, this is not possible.
If the prebuilt buntu for arm64 are available (I know that at least 17.10 is available) ,
then I’d like to modify the script for allowing that,
what do you think ?
I have added basic support for Ubuntu 18.04, you can find here:
Note: These commits may be changed in the future.
i meant: build an ubuntu 18 arm64 firmware to flash on vim board, currently, fenix scripts offer only xenial based firmware for vim, which is becoming obsolete as some says
ok, now i see it in https://github.com/numbqq/fenix/blob/master/env/setenv.sh
i will then use your version instead of the other official repo
Ok, so I deployed nick’s own version (as of yesterday) of the fenix building tool on my artful linux machine and I’ve just built,deployed,booted an ubuntu xenial + uboot2005.01 on my board,
so far so good.
Next step is connect the board to the Net using wifi, run update and dist-upgrade and recheck what does not work (if any) with this kernel 4.9
Later, I’ll replace xenial w/ artful in the process before and 4.9 with nick’s mainline
currently, fenix allows to create bionic firmware only from a bionic host (that I do not have),
my initial intention was to create an artful firmware from an artful host, by I do not really know yet what to change and in which file…
Since the vim1 pro is running this khadas firmware,
i fell it is hot, not as any other firmware i’ve used before,
is this a known issue/situation ?
I have no idea yet, I haven’t build Ubuntu artful for now, I’m not sure whether it is a bug or not…