OK, so the module compiles against the headers from the standard Debian distro (4.18-marvell). Obviously it won't work with our kernels though, so I can't test it just yet. So there's definitely something with the headers from our our kernels that is causing these compile errors...by Micky Wicky - Debian
Hi both, I'm seeing the same errors as Ayosher trying to build cryptodev (using your 4.14 kernel) on my box (NSA320S). I'll check it with the Debian Marvell kernel (Buster has a 4.18 image) when I get a chance over the next couple of weeks.by Micky Wicky - Debian
So I've tried with my Debian pendrive: umass0: HP v210w, rev 2.00/0.00, addr 4 and did run poke; sleep 120; ide reset; sleep 120; usb start; sleep 120 followed by fatload and bootm but it's made no difference at all.by Micky Wicky - uBoot
Awesome. I can confirm this patch makes kwboot work on MacOS 10.11 and on NetBSD (-current) Compiled both with clang. Nice work!by Micky Wicky - uBoot
Without poke, all 3 RC builds yield: U-Boot 2015.10-tld-2-rc3 (Mar 13 2016 - 17:12:52 -0700) ZyXEL NSA310S/320S 1/2-Bay Power Media Server SoC: Kirkwood 88F6702_A1 DRAM: 256 MiB (ECC not enabled) WARNING: Caches not enabled NAND: 128 MiB In: serial Out: serial Err: serial Net: egiga0 MV88E1318 PHY initialized on egiga0 Hit any key to stop autoboot: 0 running scan_diby Micky Wicky - uBoot
No luck... RC1 with poke: Reset IDE: Bus 0: OK Bus 1: OK --- usb start (no errors) --- sleep 60; NetBSD fails to boot with umass errors as previous RC1 without poke: Reset IDE: Bus 0: OK Bus 1: not available --> RC1 = no difference RC2 with poke: Reset IDE: Bus 0: OK Bus 1: OK --- usb start (no errors) --- sleep 60; NetBSD fails to boot with umass errors as previous RC2 without poke:Rby Micky Wicky - uBoot
Thanks for this. It makes no difference however - without the poke, ide reset only sees the first disk (and the same booting issues occur with or without poke)by Micky Wicky - uBoot
Wasn't it GPIO 33?by Micky Wicky - uBoot
Happy to test, as always.by Micky Wicky - uBoot
So I picked up an old (damaged) Seagate disk and did a couple of tests with NetBSD... HDD1 = WD, HDD2=Seagate --> no change, errors crop up the first time NetBSD tries to initialise USB HDD1 = Seagate, HDD2=WD --> NetBSD boots to a login prompt, having started all services, etc... Then the same errors happen within a couple of seconds with no intervention on my part. Note that the OS haby Micky Wicky - uBoot
Looking at my options... what can you tell me about u-boot and WD drives, Google wasn't terribly informative. Second Q - does this mean that a Seagate drive such as this one would work without any issues?by Micky Wicky - uBoot
You're not thinking about the HAL in 2001: A Space odyssey, are you? :) It's a generic acronym, The driver was first created in 2008 but received regular updates, the last one only a few weeks ago. Maybe Linux used to have a later under the same acronym, I'm not sure.by Micky Wicky - uBoot
I think this may be beyond my C skills... but let's see. The src/sys/dev/marvell/ehci_mv.c NetBSD driver is here: http://cvsweb.netbsd.org/bsdweb.cgi/~checkout~/src/sys/dev/marvell/ehci_mv.c?rev=1.5&content-type=text/plain&only_with_tag=MAIN Based on an import from 2008. It probably needs the umass driver as well (this is the one generating the error messages): http://cvsweb.by Micky Wicky - uBoot
I think that's the explanation I needed! Esp wrt the code changes in tld-2 and the WD (likely) power draw issue. So the difference between Linux and NetBSD in this is either how they initialise USB (NetBSD is asking too quickly, drawing too much power), or how the two OSes can control the power draw on SATA (if that is possible...) I'll see what I can make out of the USB stuff in Nby Micky Wicky - uBoot
QuoteSo did it boot all the way to the Debian prompt here? Yes it did QuoteHave you tried a different USB flash drive? what brand are you using? umass0 is a Kingston DataTraveler 2.0, rev 2.00/2.00, addr 3 16GB with NetBSD installation on it umass1 is a (sloooooooooowwwww) HP v210w, rev 2.00/0.00, addr 4 16GB with Debian installation on it I've got a 1GB stick lying aroundby Micky Wicky - uBoot
With 2 HHDs, doing run poke; ide reset; sleep 60; usb start Launch Debian (the macro scan_disk includes "ide reset") - looks OK: root@debian:~# dmesg | grep -i SATA [ 2.626551] sata_mv f1080000.sata: version 1.28 [ 2.626885] sata_mv f1080000.sata: slots 32 ports 2 [ 2.703914] scsi host0: sata_mv [ 2.723822] scsi host1: sata_mv [ 2.727368] ata1: SATA max UDMA/1by Micky Wicky - uBoot
I know the feeling - I normally use a RPI-2, but it runs NetBSD! I'll report if I ever get around to debugging kwboot properly.by Micky Wicky - uBoot
> No - no - yes - OK - OK. ******* HDD2 was not initialized? But boot OK? Correct > Yes - no - yes - OK - OK. ****** HDD2 was initilized, and boot OK? Correctby Micky Wicky - uBoot
You're right, I just had to git pull and it builds using clang 7.0 on 10.11.1 Running it returns the exact same error as under NetBSD 7.0: kwboot -B 115200 -t -p -b uboot.2015.10-tld-1.nsa310s.mtd0.kwb /dev/tty.SLAB_USBtoUART Sending boot message. Please reboot the target...\ Sending boot image... 0 % [+xmodem: Protocol error Did you see this as well? I tried to track it downby Micky Wicky - uBoot
Haven't figured out how to do tables in here.... So I tried different combinations of poke/no poke, and whether HDDs are connected, in Debian and NetBSD (both using USB rootfs): this is doing no more than the poke and run bootcmd_default (for Debian) and usb start; fatload; bootm (for NetBSD) Poke - HDD1 - HDD2 - Debian Result - NetBSD Result No - yes - yes - FAIL - FAIL No - yes - noby Micky Wicky - uBoot
So, having no disk in slot 1 and a disk in slot 2. With your updated envs (saved) poke and boot - Debian boots successfully NSA3x0S> run debian_boot starting USB... USB0: USB EHCI 1.00 scanning bus 0 for devices... 4 USB Device(s) found scanning usb for storage devices... Use USB retry period from the environment: 45 second(s) Use USB retry period from the environment:by Micky Wicky - uBoot
No problem on OS X 10.6 (32-bit) and 10.11 in my case, I got the drivers from https://www.silabs.com/products/mcu/Pages/USBtoUARTBridgeVCPDrivers.aspx#mac I don't recall any issues with the Kexts. Re: compiling Linux code on Mac, it's a good starting point, but kwboot won't work without some changes. I briefly looked into it, it doesn't compile in the first place. Maybeby Micky Wicky - uBoot
Yeah, on Mac you need to install the CP2102 kexts from the web (I think I went to the vendor website, it was pretty obvious at the time so didn't think twice about it).by Micky Wicky - uBoot
So test #3 fails (I can't try the other 2 as I only have the 2 WD disks) The timeouts do not occur if there is no HDD present as we've seen earlier. U-Boot 2015.10tld-2 (Feb 23 2016 - 02:17:09 -0800) ZyXEL NSA310S/320S 1/2-Bay Power Media Server SoC: Kirkwood 88F6702_A1 DRAM: 256 MiB (ECC not enabled) WARNING: Caches not enabled NAND: 128 MiB In: serial Out: seriby Micky Wicky - uBoot
It'll be a couple of days before I can test on my side, I'll let you know how it goes.by Micky Wicky - uBoot
I mounted the USB on another ARM Linux device to read it: odroid:/home/michael$ sudo e2label /dev/sda1 rootfs odroid:/home/michael$ sudo mount /dev/sda1 /mnt/ odroid:/home/michael$ sudo chrootby Micky Wicky - uBoot
The uImage had the DTB embedded, I followed your guide to do this. Not sure about the initrd, I'll have to check tomorrow. Shouldn't be the issue though. The partition is definitely labeled rootfs. Otherwise it wouldn't work on tld-1, right? The fstab refers to the root partition as rootfs. I'll post the detail tomorrow. I'm really not convinced that the problem is OS-by Micky Wicky - uBoot
There you go, see full log below. Worth noting that I have 3 USB drives that I've used to boot this NSA320s, all 3 (1 NetBSD, 2 Debian Jessie) from different manufacturers and sizes, all 3 have booted successfully from tld-1 in the last 48 hours. But with tld-2 all of them show the same issue (regardless of OS). As a freebie I've run ext4ls - so u-boot can absolutely read from USB, buby Micky Wicky - uBoot
So, using tld-2... - Pluged in the Debian USB flash rootfs, and 2 HDDs. - Cold started with kwboot, interrupted serial console - setenv bootcmd_default 'run scan_disk; run set_bootargs; run bootcmd_exec' - run bootcmd_default --> Debian starts but fails as earlier with Gave up waiting for root device. Common problems: - Boot args (cat /proc/cmdline) - Check roby Micky Wicky - uBoot
All of the below was done from a cold start. 1. As you expected, I can boot from USB with tld-2 as long as the 2 HDDs are physically disconnected. 2. setting usb_ready_try to 45 - no change: NSA3x0S> setenv usb_ready_retry 45 NSA3x0S> boot starting USB... USB0: USB EHCI 1.00 scanning bus 0 for devices... 3 USB Device(s) found scanning usb for storage devices... Use Uby Micky Wicky - uBoot