Welcome! Log In Create A New Profile

Advanced

Debian on Action Tech SG200 and MI424WR-rev-I

Posted by veriqster 
Re: Debian on Action Tech SG200 and MI424WR-rev-I
July 31, 2021 11:34PM
bodhi,

OH, now the driver is naughty. I updated compat-wireless package and there is an oops message which I can't remove.
[   18.923140] mvneta f1074000.ethernet eth1: link up
[   18.939346] Unable to handle kernel NULL pointer dereference at virtual address 00000006
[   18.947478] pgd = c59c0000
[   18.950198] [00000006] *pgd=05c7d831, *pte=00000000, *ppte=00000000
[   18.956526] Internal error: Oops: 817 [#1] ARM
[   18.960989] Modules linked in: rtl8192ce rtl_pci rtl8192c_common pppoe ppp_async iptable_nat rtlwifi pppox ppp_generic nf_nat_ipv4 nf_conntrack_ipv6 nf_conntrack_ipv4 mac80211 ipt_MASQUERADE cfg80211 xt_time xt_tcpudp xt_tcpmss xt_statistic xt_state xt_recent xt_nat xt_multiport xt_mark xt_mac xt_limit xt_length xt_id xt_hl xt_helper xt_ecn xt_dscp xt_conntrack xt_connmark xt_connlimit xt_connbytes xt_comment xt_TCPMSS xt_REDIRECT xt_LOG xt_HL xt_DSCP xt_CT xt_CLASSIFY usbhid ums_usbat ums_sddr55 ums_sddr09 ums_karma ums_jumpshot ums_isd200 ums_freecom ums_datafab ums_cypress ums_alauda slhc nf_nat_irc nf_nat_ftp nf_nat nf_defrag_ipv6 nf_defrag_ipv4 nf_conntrack_rtcache nf_conntrack_irc nf_conntrack_ftp nf_conntrack iptable_raw iptable_mangle iptable_filter ipt_REJECT ipt_ECN ip_tables hid_generic crc_ccitt compat fuse hid evdev input_core ledtrig_usbdev ledtrig_oneshot xt_LED ledtrig_morse ledtrig_heartbeat ledtrig_gpio xt_set ip_set_list_set ip_set_hash_netiface ip_set_hash_netport ip_set_hash_netnet ip_set_hash_net ip_set_hash_netportnet ip_set_hash_ipportnet ip_set_hash_ipportip ip_set_hash_ipport ip_set_hash_ip ip_set_bitmap_port ip_set_bitmap_ipmac ip_set_bitmap_ip ip_set nfnetlink ip6t_REJECT ip6table_raw ip6table_mangle ip6table_filter ip6_tables x_tables msdos vfat fat isofs nls_utf8 nls_iso8859_1 nls_cp437 ipv6 arc4 crypto_blkcipher ledtrig_transient exfat
[   19.083783] CPU: 0 PID: 2539 Comm: hostapd Not tainted 3.14.79 #1
[   19.089905] task: c71ff420 ti: c5cb4000 task.ti: c5cb4000
[   19.095347] PC is at rtl_get_tcb_desc(rtlwifi/base.c)+0x68/0x6a8 [rtlwifi]
tcb_desc->hw_rate = txrate->hw_value;
[   19.100866] LR is at rtl92ce_tx_fill_desc(rtlwifi/rtl8192ce/trx.c)+0x17c/0x634 [rtl8192ce]
[   19.106989] pc : [<bf36a350>]    lr : [<bf3a6218>]    psr: 80000093
[   19.106989] sp : c5cb5a18  ip : c5cb5a48  fp : c5cb5a44
[   19.118518] r10: c7391720  r9 : 00000080  r8 : c58be7a0
[   19.123766] r7 : 00000000  r6 : c5c8def8  r5 : c7390b00  r4 : 00000000
[   19.130322] r3 : 00000000  r2 : 00000000  r1 : bf3740b8  r0 : c7390b00
[   19.136879] Flags: Nzcv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  Segment user
[   19.144134] Control: 0005397f  Table: 059c0000  DAC: 00000015
[   19.149905] Process hostapd (pid: 2539, stack limit = 0xc5cb41b8)
[   19.156025] Stack: (0xc5cb5a18 to 0xc5cb6000)
[   19.160404] 5a00:                                                       c5c8dee0 c8805000
[   19.168622] 5a20: 00000000 c7391720 c58be7a0 00000000 00000000 00000000 c5cb5a94 c5cb5a48
[   19.176842] 5a40: bf3a6218 bf36a2f8 00000000 00000000 c5ed18a4 00000000 00000010 c5ed1b20
[   19.185060] 5a60: 058be7a0 00000080 c5ed1400 c7391720 c5c8dee0 c739b6fc c7390b00 c58be7a0
[   19.193279] 5a80: 00009e1c c8805000 c5cb5afc c5cb5a98 bf3982d4 bf3a60ac c5c8def8 00000000
[   19.201498] 5aa0: c5c8dee0 00000004 00000000 00000001 00000200 00000000 00000000 00000000
[   19.209716] 5ac0: 00000000 00000013 c7391758 01040200 c5cb5afc c7391720 000c8340 c7390b00
[   19.217936] 5ae0: c5ed1b28 c5ed1b20 c7391758 00000200 c5cb5b3c c5cb5b00 bf36df8c bf397fa4
[   19.226155] 5b00: 00098583 c7391720 c5ed140c c5ed1400 c5cb5b34 00000001 c5ed1000 00000100
[   19.234373] 5b20: c5ed189c 00000356 c5cb5bf4 c7390b00 c5cb5b5c c5cb5b40 bf2e4768 bf36de8c
[   19.242593] 5b40: c5ed189c c5ed140c c5ed1000 c5ed140c c5cb5b9c c5cb5b60 bf2fc128 bf2e46b4
[   19.250813] 5b60: c7390140 00000003 c5ed1000 c5ed1400 c5cb5b9c c5ed1428 c5ed1408 c7390000
[   19.259032] 5b80: c5ed1000 c5cb5ba0 c7390140 c71cd400 c5cb5c6c c5cb5ba0 bf2bce70 bf2fbebc
[   19.267251] 5ba0: bf374830 00000001 0000099e 00000000 c5c60620 c5c60660 c5c606fc c5c6070c
[   19.275469] 5bc0: c5c6071c 00000000 0000003a 00000061 0000000a 0000000a 0000000a 00000000
[   19.283687] 5be0: 00000064 00000002 c5c606d8 00000007 00000000 00000000 00000000 00000000
[   19.291904] 5c00: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[   19.300122] 5c20: 8e880000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[   19.308342] 5c40: c5cb5c6c bf2d13c0 bf2c8c14 c5c60600 c5c8d820 00000000 c5c60610 c71cd400
[   19.316560] 5c60: c5cb5ccc c5cb5c70 c02beec0 bf2bcb04 bf2c877c 00000000 c71f32a0 60ffaca6
[   19.324780] 5c80: 000009eb c5c60600 c5c60610 c5c60614 c71cd400 c7390000 c5ed1000 c786e000
[   19.333000] 5ca0: c028f1d0 c5c60600 c5c8d820 c02bec00 00000128 00000000 00000000 00000000
[   19.341218] 5cc0: c5cb5cec c5cb5cd0 c02be38c c02bec10 60000013 c044002c c5c8d820 c5c8d820
[   19.349437] 5ce0: c5cb5d04 c5cb5cf0 c02bebec c02be344 c786e000 c714f400 c5cb5d34 c5cb5d08
[   19.357655] 5d00: c02bdcfc c02bebd4 c5cb5e68 7fffffff 00000000 00000000 c5cb5f5c c5cb5dd8
[   19.365873] 5d20: c714f400 c5c8d820 c5cb5

8086k
Re: Debian on Action Tech SG200 and MI424WR-rev-I
August 26, 2021 12:39AM
Good news:
I have ported sg200's uboot to F660 and HGG420N. Now everything works fine. Everyone can compile it.
NOTE:
You must use 32bit linux to compile it

Github repo: https://github.com/msdos03/avanta-uboot-f660
https://github.com/msdos03/avanta-uboot-hgg420n

8086k



Edited 1 time(s). Last edit at 10/01/2021 03:52AM by 8086k.
Re: Debian on Action Tech SG200 and MI424WR-rev-I
October 27, 2021 10:26AM
@bodhi

I have done the patch to make mvneta support 88f6560. Your DTS need to be changed to use this patch, so I will post my F660 DTS. Please add my patch to your debian work.
I also proved that kernel DSA driver doesn't work with 88f6560 internal switch, so I dumped mvsw61xx from older openwrt and updated some newer features to it according to the commits. The driver and Kconfig/Makefile patches will be posted too.

All my work are done for kernel 5.10.75 . You may need to change it a bit.

This is a part of new log

[    1.230906] mvsw61xx 8.mvsw6171: Found MV88F6560 built-in at f1072004.mdio-bus-mii:08
[    1.238805] mvsw61xx 8.mvsw6171: Using indirect addressing
[    1.280684] Freeing unused kernel memory: 10240K
[    1.285943] Run /init as init process
[    1.673744] init: Console is alive
[    1.688883] kmodloader: loading kernel modules from /etc/modules-boot.d/*
[    1.724951] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
[    1.734181] ehci-fsl: Freescale EHCI Host controller driver
[    1.742897] ehci-orion: EHCI orion driver
[    1.747151] orion-ehci f1050000.ehci: EHCI Host Controller
[    1.752709] orion-ehci f1050000.ehci: new USB bus registered, assigned bus number 1
[    1.760570] orion-ehci f1050000.ehci: irq 26, io mem 0xf1050000
[    1.793361] orion-ehci f1050000.ehci: USB 2.0 started, EHCI 1.00
[    1.800053] hub 1-0:1.0: USB hub found
[    1.804555] hub 1-0:1.0: 1 port detected
[    1.811608] ehci-platform: EHCI generic platform driver
[    1.825381] usbcore: registered new interface driver usb-storage
[    1.834291] kmodloader: done loading kernel modules from /etc/modules-boot.d/*
[    1.852006] init: - preinit -
[    2.046959] random: jshn: uninitialized urandom read (4 bytes read)
[    2.098701] random: jshn: uninitialized urandom read (4 bytes read)
[    2.125950] random: jshn: uninitialized urandom read (4 bytes read)
[    2.320836] mvneta f1070000.ethernet eth0: configuring for fixed/gmii link mode
[    2.329583] mvneta f1070000.ethernet eth0: Link is Up - 1Gbps/Full - flow control off
[    2.342898] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
Press the [f] key and hit [enter] to enter failsafe mode
Press the [1], [2], [3] or [4] key and hit [enter] to select the debug level
[    4.123332] random: fast init done
[    6.456132] mvneta f1070000.ethernet eth0: Link is Down
[    6.485345] procd: - early -
[    7.025707] procd: - ubus -
[    7.034205] urandom_read: 6 callbacks suppressed
[    7.034212] random: ubusd: uninitialized urandom read (4 bytes read)
[    7.082425] procd: - init -
Please press Enter to activate this console.
[    7.549173] kmodloader: loading kernel modules from /etc/modules.d/*
[    7.727844] urngd: v1.0.2 started.
[    7.809338] random: crng init done
[    7.874519] Loading modules backported from Linux version v5.10.68-0-g4d8524048a35
[    7.882137] Backport generated by backports.git v5.10.68-1-0-ga4f9ba32
[    7.963866] xt_time: kernel timezone is -0000
[    8.070164] PPP generic driver version 2.4.2
[    8.085118] NET: Registered protocol family 24
[    8.163264] pci 0000:00:02.0: enabling device (0140 -> 0143)
[    8.169165] rtl8192ce: Chip Version ID: B_CHIP_92C
[    8.183389] rtl8192ce: Using firmware rtlwifi/rtl8192cfw.bin
[    8.198828] rtlwifi: rtlwifi: wireless switch is on
[    8.249092] kmodloader: done loading kernel modules from /etc/modules.d/*
[   48.243880] mvneta f1070000.ethernet eth0: configuring for fixed/gmii link mode
[   48.251433] mvneta f1070000.ethernet eth0: Link is Up - 1Gbps/Full - flow control off
[   48.331021] br-lan: port 1(eth0) entered blocking state
[   48.336325] br-lan: port 1(eth0) entered disabled state
[   48.341849] device eth0 entered promiscuous mode
[   48.380845] br-lan: port 1(eth0) entered blocking state
[   48.386148] br-lan: port 1(eth0) entered forwarding state
[   49.283407] IPv6: ADDRCONF(NETDEV_CHANGE): br-lan: link becomes ready

8086k
Attachments:
open | download - avanta-net-fix.tar (77 KB)
Re: Debian on Action Tech SG200 and MI424WR-rev-I
October 27, 2021 03:51PM
8086k,

> I also proved that kernel DSA driver doesn't work
> with 88f6560 internal switch, so I dumped mvsw61xx
> from older openwrt and updated some newer features
> to it according to the commits. The driver and
> Kconfig/Makefile patches will be posted too.

No, you have not proven that the DSA driver does not work for this SoC. What you did is a work-around to bring back the old switch driver.

Sorry, this solution is not good enough for me to include it in the kernel. The right solution is, as a first step, solving the IRQ conflicts, to bring up the ethernet ports. And then the DSA driver would take over and initialize the DSA switch properly.

But please feel free to post whatever you'd like, in case others want to use it in their own boxes.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)
Re: Debian on Action Tech SG200 and MI424WR-rev-I
October 28, 2021 09:28AM
bodhi,

You don't understand me, please have a look at the attachment. I have fixed the new MVNETA driver of 5.10 kernel(all the issues are fixed, including IRQ and GMII mode), not bringing back the old one, the patch is in the attachment. My MVNETA patch-work is tested good.

And I have also tested using my fixed mvneta driver with DSA driver using several different configs, I changed the identify code in DSA driver but it still doesn't work.

mvsw61xx is a switch driver not included in kernel. I only bring this back. And this driver works very well with no errors.

So I think you should include it in the kernel.

8086k
Re: Debian on Action Tech SG200 and MI424WR-rev-I
October 28, 2021 01:56PM
8086k,

If your patch for the IRQ works with the latest kernel, then the next step is the DTS nodes for the DSA switch.

The DSA driver was based from the old switch driver. So if the old switch driver works then the DSA driver should work. Just a matter of correct configuration is needed in the DTS.

See the example in the Linksys EA4500 DTS.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)
Re: Debian on Action Tech SG200 and MI424WR-rev-I
October 29, 2021 04:51AM
bodhi,

My next step is to solve realtek driver issue. The wireless is more important than using dsa switch driver. Fixing that is almost useless.

8086k
Re: Debian on Action Tech SG200 and MI424WR-rev-I
October 29, 2021 04:08PM
@8086k,

OK. Sounds like we don't have the same goal, and are not in the same project.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)
Aruba AP-135
January 27, 2022 07:57PM
Hey, not to drag up dead topics, but I recently picked up an Aruba AP-135 in a thrift store, so if you'd like to see if we can get your Debian to run on it, I'm happy to test for you.
Re: Aruba AP-135
January 27, 2022 09:21PM
bkallus,

> Hey, not to drag up dead topics,

Not at all! I'm just too busy to revisit this. I hope someday I'll have some free time to look at the Ethernet IRQ conflicts problem. Currently the MI424WR is running fine with Wifi network connection.

> but I recently
> picked up an Aruba AP-135 in a thrift store, so if
> you'd like to see if we can get your Debian to run
> on it, I'm happy to test for you.

Sure, go ahead and use the steps I posted in this thread to see if you can boot Debian on USB drive.

And let me see. I'll post the DTB for this box.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)



Edited 1 time(s). Last edit at 01/27/2022 10:53PM by bodhi.
Re: Aruba AP-135
January 28, 2022 09:11PM
bkallus,

The initial DTB for Aruba was attached to this post:

https://forum.doozan.com/read.php?2,78499,120849#msg-120849

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)
Re: Aruba AP-135
January 30, 2022 03:49PM
Hey there. I picked up an AP-134 (identical to the 135, but external antennas).

Aruba's U-boot does not include a `bootm`, but only an `apboot`, which also checks what you boot against a certificate stored in flash (16MiB MXIC SPI NOR). A copy of my device's flash is here: https://downloads.laboratoryb.org/dumps/AP135/

Its U-boot also comes with md and mm, so one way in would be to overwrite the RSA cert (trivial). However, I've been really wanting to get kwboot to work (since then, a u-boot replacement would be trivial for end-users). For some reason, though, the kwboot handshake is never accepted.

Reasons I checked:
- The external console is ttyS0 for sure, based on the strings in the dump.

Other reasons:
- Maybe the MA3232C (https://www.ti.com/lit/ds/symlink/max3232.pdf?ts=1643550007405) which level-shifts from the console to the UART is in the way. There is what appears to be a set of UART pins; I can test those and see if they work.
- Maybe there's something in the BootROM that messes this up, like a jumper (like those required for Dove boards). Marvell never opened the functional specifications sheet for the Avanta family, so we may never know: https://www.kernel.org/doc/html/latest/arm/marvell.html#avanta-family
Re: Aruba AP-135
January 30, 2022 04:41PM
hurricos,


> - The external console is ttyS0 for sure, based on the strings in the dump.

Which dump did you look at? the part where mtd0 (i.e. u-boot), or the entire dump? It could be in the kernel part too.

> - Maybe there's something in the BootROM that
> messes this up, like a jumper (like those required
> for Dove boards).

Not likely the reason, since I can kwboot the MI424WR-rev-I. So we are sure that this BootROM does not have the same quirk as the Dove SoC. Unless a different BootROM is on the Aruba vs the MI424 (unlikely).

If regular serial console works on UART0 , then the box is set so that the BootROM will poll that UART0 port. Unless the vendor did some thing in the HW to make BootROM polling UART1.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)
Re: Aruba AP-135
January 31, 2022 04:11PM
You did mention that electrical interference may screw things up -- it's possible that
1) as I mentinoed, the MA3232C chip might be interfering somewhat, or
2) (and this would be even cleverer on Aruba's part), that the MA3232C isn't reliably fully initialized by the time the device is fully powered on.

A solution to #2 would be to find a clean way to force a soft reset which would leave the MA3232C fully powered and in working state.

> Which dump did you look at? the part where mtd0 (i.e. u-boot), or the entire dump? It could be in the kernel part too.

I looked through the whole dump -- though I never went through and did e.g. a binwalk -Me to pull apart all the little bits -- I figured since the partitions were JFFS2, something as small as 'ttyS1' or even S1 would show up clear as day.

I don't think the second internal-console-looking thing is actually wired to the UART side of ttyS0, but I will test with my logic analyzer tonight.
Re: Aruba AP-135
January 31, 2022 04:54PM
> I looked through the whole dump -- though I never
> went through and did e.g. a binwalk -Me to pull
> apart all the little bits -- I figured since the
> partitions were JFFS2, something as small as
> 'ttyS1' or even S1 would show up clear as day.

The 1st MB is u-boot. If the ttyS0 is configured in, then it should be there,

>
> I don't think the second internal-console-looking
> thing is actually wired to the UART side of ttyS0,
> but I will test with my logic analyzer tonight.

Yeah, very unlikely. But if you boot up and see nothing in serial console, that could be it.

OTOH, if the input/output is available, then you'll know the port is the one, but that port could be UART0 or UART1.

Usually, this info is in the Reference Manual, but we don't have that. However, we do have GPL source for the MI424. This u-boot source might have Aruba patch (I did not look for Aruba so don't know yet).

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)
Re: Aruba AP-135
February 03, 2022 04:04PM
The AP-135 doesn't have a USB port, so I can't test with your rootfs.
My apboot does have tftpboot, so I should be able to get something working with that. Unfortunately, I see no way to set a static IP, so I'll be wrangling with apboot's dhcp client for a while before tftpboot will work.
Re: Aruba AP-135
February 03, 2022 04:10PM
bkallus,

> The AP-135 doesn't have a USB port, so I can't
> test with your rootfs.

> My apboot does have tftpboot, so I should be able
> to get something working with that.

What is apboot? and please remind me if we have specs for the AP-135 somewhere on this thread.

> Unfortunately,
> I see no way to set a static IP, so I'll be
> wrangling with apboot's dhcp client for a while
> before tftpboot will work.

Can you connect and run serial console to interact with u-boot?


Update:

Found specs on http://en.techinfodepot.shoutwiki.com/wiki/Aruba_Networks_AP-135

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)



Edited 1 time(s). Last edit at 02/03/2022 04:14PM by bodhi.
Re: Aruba AP-135
February 03, 2022 04:18PM
Looks like the this Aruba-135 is more appropriate running OpenWrt on flash (no other storage device). So what we can do here is to get it boot with the DTS. And then either:

- Run Debian on nfs rootfs
or
- Run OpenWrt on flash.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)
Re: Aruba AP-135
February 04, 2022 08:04AM
bkallus,

Look at my repo. It is openwrt snapshot and still updating regularly. In my repo, I fixed mvneta and orion nand driver. Only need a new device tree to add your device support.
https://github.com/msdos03/openwrt-avanta
Re: Aruba AP-135
February 04, 2022 10:02AM
> What is apboot?
apboot is Aruba's U-boot.

> Can you connect and run serial console to interact with u-boot?
Yes. According to this, I can just setenv serverip and ipaddr and avoid using DHCP like that. I'll try that when I get back home in a few hours.

> Look at my repo. It is openwrt snapshot and still updating regularly.
Very cool; I'll check this out. I'll try sticking in the MI424WR-I dts too and see if things work.
Re: Aruba AP-135
February 04, 2022 08:14PM
bkallus,
There is a repo of the GPL code for aruba there
https://github.com/shalzz/aruba-ap-310
If it is usable, then cracking uboot will be so easy. Try to compile it and test on your device. The apboot included there is based on uboot for 88f6560, so it's likely to support your device. If it works, we can remove apboot command and add back standard bootm command.

Another option is to use 'go' command in uboot to start zimage. Kernel commandlines are written in dts.
Re: Aruba AP-135
February 05, 2022 06:16PM
These are the commands available from within apboot:
apboot> help
?              - alias for 'help'
boot           - boot the OS image
clear          - clear the OS image or other information
dhcp           - invoke DHCP client to obtain IP/boot params
factory_reset  - reset to factory defaults
help           - print online help
mfginfo        - show manufacturing info
osinfo         - show the OS image version(s)
ping           - send ICMP ECHO_REQUEST to network host
printenv       - print environment variables
purgeenv       - restore default environment variables
reset          - Perform RESET of the CPU
saveenv        - save environment variables to persistent storage
setenv         - set environment variables
tftpboot       - boot image via network using TFTP protocol
upgrade        - upgrade the APBoot or OS image
version        - display version
These are the default environment variables:
apboot> printenv
bootdelay=2
baudrate=9600
autoload=n
boardname=Arran
bootcmd=boot ap
autostart=yes
bootfile=armv5te.ari
ethaddr=6c:f3:7f:cc:ce:8e
name=BL007-FF-009A-WD-A
group=BL007-FF-Franklin Hall
master=aruba-master.indiana.edu
ip6prefix=64
servername=aruba-master
a_antenna=0
g_antenna=0
usb_type=0
uplink_vlan=0
auto_prov_id=0
is_rmp_enable=0
priority_ethernet=0
priority_cellular=0
4g_usb_type=0
cert_cap=0
mesh_role=0
installation=0
mesh_sae=0
num_ipsec_retry=85
start_type=cold_start
stdin=serial
stdout=serial
stderr=serial
ethact=eth0
This router came from a thrift store near Indiana University, so that's why there are references to IU in there.
tftpboot seems like our best bet here, but it's pretty nonstandard. Running tftpboot starts up a DHCP client and acquires an IP, then asks for aruba-master (from the servername variable) over DNS. I have never set up a DNS server before, so I still haven't gotten that query to succeed. I wouldn't be surprised if it expects the tftpboot images to be signed just like the images on its flash.

If I were able to compile those sources, how would I go about running it on the AP-135? Thanks.
Re: Aruba AP-135
February 05, 2022 07:51PM
bkallus,
Use kwboot to boot the image from UART.
I think Indiana University has removed some commands from u-boot. You may look for others who has an ap135 and ask for a copy of U-Boot(if it supports bootm). This is a better idea. I don't want to port any U-Boot again because it's very hard and need a lot of tests on device directly.
I saw someone said U-Boot is in spi flash. So the first thing you should do is find where is U-Boot stored and try to access it in U-Boot as well as in OS.
Re: Aruba AP-135
February 06, 2022 08:08AM
For whatever reason, I can't get the kwboot handshake to work, just like hurricos. hurricos's uboot also has no bootm, so i bet that all the 134s and 135s dont. My U-Boot has no mm or md, (hurricos's does) so that suggests that maybe IU did mess around with U-Boot. As for accessing stuff within the OS, the serial shell user is extremely locked down; no permission to do anything. All I have been able to do so far is leak environment variables.

EDIT: I don't think IU locked anything down in U-Boot; I actually do have md and mm. They're just not listed in the help message.



Edited 1 time(s). Last edit at 02/06/2022 09:00AM by bkallus.
Re: Aruba AP-135
February 06, 2022 04:20PM
bkallus,

Try this kwboot binary I recently built and attched to this post:

https://forum.doozan.com/read.php?3,129444,129919#msg-129919

If the new kwboot still does not work, try different timing such as: start kwboot and power up, power up and then start kwboot,.... The MI424 is straight forward, kwboot starts handshake right away.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)
Re: Aruba AP-135
February 06, 2022 07:24PM
bkallus,
This U-Boot must be replaced. There is no 'nand' or 'sf' command to access flash memory. It looks like the only way is to use that GPL code. Kwboot is in u-boot-tools package. This is the only working Kwboot for me. Don't use too old linux OS because their u-boot-tools don't include Kwboot.
Re: Aruba AP-135
February 06, 2022 08:40PM
> Kwboot
> is in u-boot-tools package. This is the only
> working Kwboot for me. Don't use too old linux OS
> because their u-boot-tools don't include Kwboot.

You should try the latest kwboot that I posted. It is practically a new kwboot, more modern, and very robust. In the past, some boxes cannot be booted easily because old bugs in old kwboot (which is what in all current distros).

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)
Re: Aruba AP-135
February 06, 2022 09:34PM
I have been using a kwboot that I built from the newest U-Boot source. I am able to successfully kwboot the MI424WR, so I know the binary is good. It just doesnt work on the Ap-135.

Do you have any tips for building those GPL sources? I'm having trouble finding the armeb-linux-gnu-uClibc-0.9.30 toolchain. Can I use the armeb-linux-gnueabi toolchain instead?
Re: Aruba AP-135
February 06, 2022 10:38PM
bkallus,

> Do you have any tips for building those GPL
> sources? I'm having trouble finding the
> armeb-linux-gnu-uClibc-0.9.30 toolchain. Can I use
> the armeb-linux-gnueabi toolchain instead?

GPL source should be built using the toolchain that's included in the GPL source packages. Modern toolchain will be too differrent and usually no longer have the old stuff. It's always a struggle to try to build very old source using modern toolchain, so I would not try that.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)
Re: Aruba AP-135
February 07, 2022 06:33AM
As far as I can tell, the toolchain is not included in the code from github. I'm not sure where I can get the old toolchain required, with the uClibc ABI. Do you think I can use a gnueabi version instead, given that I only need to compile U-Boot (which I assume won't need to dynamically link to anything)?
Author:

Your Email:


Subject:


Spam prevention:
Please, enter the code that you see below in the input field. This is for blocking bots that try to post this form automatically. If the code is hard to read, then just try to guess it right. If you enter the wrong code, a new image is created and you get another chance to enter it right.
Message: