New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
FS#3800 - ath79: trunk is bricking old ath9k devices #8831
Comments
Greybeard: Not exactly sure if this is relevant but I haven't been able to use trunk (ath79) on an tp-link wr842v1 for the last month. Every time I try and flash it somewhere the process fails and I have to do a failsafe tftp restore of a factory image to get it back again. I also have a tplink c7v5 but not game to flash it with a current trunk as it is my primary router. |
mgondium: I found that the last trunk version that I upgraded successfully to was r15661-cbedb5de7, so it has to be some commit since then. OR, the upgrade process in that one was borked and bricked the new one. EDIT: just tested another device, r15661 to So it's trunk, sometime since r15661. |
GKaURTbV: Can you provide the links to the firmware images you used? |
None: I experienced problem after this build: Though it seems following commit is breaking trunk for old devices: More details in this forum thread Hope this helps ;) |
walterav1984: Quotes: Do you all mean 21.02-rc1 instead of 20.02-x? This patch from Rafał Miłecki fixed ath79 mtd rootfs detection on other devices in the snapshot of 21.02 since rc1 was broken: Maybe worth a try: |
mgondium: My offending image was made via buildroot from trunk. BE AWARE THAT THIS WILL BRICK THE DEVICE, DO NOT FLASH IT. https://filebin.net/jlzei8w1jjujn6ws/openwrt-ath79-generic-tplink_tl-wr1043nd-v1-squashfs-sysupgrade.bin My current working images were build after "git checkout openwrt-21.02", is this RC1? @walter Sonius |
None: Hi, I tried to flash image openwrt-ath79-generic-tplink_tl-wr1043nd-v1-squashfs-factory.bin - r16723-7fea9d9f5d - on top of factory firmware after build https://buildbot.openwrt.org/master/images/#/builders/1/builds/111 No luck. Router does not boot at all. I was able flash back last factory firmware via tftp, though. It seems, following commit did not fix issue for TP-Link 1043nd-v1. |
GKaURTbV: Hi, About the TP-Link TL-WR1043ND v1 - I am not really sure what are you talking about. I have one. OpenWrt 21.02.0-rc1 successfully runs on that model, and as a matter of fact, I have been running the rc1 build ever since I flashed it on the device, more than 2 weeks now. You can get this firmware by going to the firmware selector, selecting 21.02.0-rc1 and searching for 1043 - link: From the results - I chose "TP-Link TL-WR1043N/ND v1". This is the build info I get in the status page on the router: And if any help - I am also attaching a boot log. Keep in mind that with OpenWrt 19.x, I switched my 1043ND to ath79, although the ar71xx builds were still supported. With OpenWrt 21.x this changes - we only have the ath79 to move forward for the 1043ND. Hope this helps. |
GKaURTbV: As for the trunk problems - see the comments on this pull request: There is a link at the end with working firmwares for CPE and WBS devices. The fixes are not merged yet, though. |
None: @GKaURTbV
In a few minutes ago I flashed back r16383-ec6293febc nightly/daily of master/trunk with success using same method when I tried to flash r16723-7fea9d9f5d with no success. This is the last build of master/trunk which is working OK. Hope fix will come soon into trunk branch, so I can install the new patches for WiFi CVEs.
Sat May 15 13:59:17 2021 kern.notice kernel: [ 0.000000] Linux version 5.4.109 (builder@buildhost) (gcc version 8.4.0 (OpenWrt GCC 8.4.0 r16383-ec6293febc)) #0 Fri A
pr 2 14:45:28 2021
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] printk: bootconsole [early0] enabled
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] CPU0 revision is: 00019374 (MIPS 24Kc)
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] MIPS: machine is TP-Link TL-WR1043ND v1
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] SoC: Atheros AR9132 rev 2
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Initrd not found or empty - disabling initrd
Sat May 15 13:59:17 2021 kern.warn kernel: [ 0.000000] Primary instruction cache 64kB, VIPT, 4-way, linesize 32 bytes.
Sat May 15 13:59:17 2021 kern.warn kernel: [ 0.000000] Primary data cache 32kB, 4-way, VIPT, cache aliases, linesize 32 bytes
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Zone ranges:
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Normal [mem 0x0000000000000000-0x0000000001ffffff]
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Movable zone start for each node
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Early memory node ranges
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] node 0: [mem 0x0000000000000000-0x0000000001ffffff]
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Initmem setup node 0 [mem 0x0000000000000000-0x0000000001ffffff]
Sat May 15 13:59:17 2021 kern.debug kernel: [ 0.000000] On node 0 totalpages: 8192
Sat May 15 13:59:17 2021 kern.debug kernel: [ 0.000000] Normal zone: 72 pages used for memmap
Sat May 15 13:59:17 2021 kern.debug kernel: [ 0.000000] Normal zone: 0 pages reserved
Sat May 15 13:59:17 2021 kern.debug kernel: [ 0.000000] Normal zone: 8192 pages, LIFO batch:0
Sat May 15 13:59:17 2021 kern.debug kernel: [ 0.000000] pcpu-alloc: s0 r0 d32768 u32768 alloc=1*32768
Sat May 15 13:59:17 2021 kern.debug kernel: [ 0.000000] pcpu-alloc: [0] 0
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Built 1 zonelists, mobility grouping on. Total pages: 8120
Sat May 15 13:59:17 2021 kern.notice kernel: [ 0.000000] Kernel command line: console=ttyS0,115200 rootfstype=squashfs,jffs2
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Dentry cache hash table entries: 4096 (order: 2, 16384 bytes, linear)
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Inode-cache hash table entries: 2048 (order: 1, 8192 bytes, linear)
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Writing ErrCtl register=00000000
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Readback ErrCtl register=00000000
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] mem auto-init: stack:off, heap alloc:off, heap free:off
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] Memory: 24376K/32768K available (5229K kernel code, 190K rwdata, 1192K rodata, 1188K init, 205K bss, 8392K res
erved, 0K cma-reserved)
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] SLUB: HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] NR_IRQS: 51
Sat May 15 13:59:17 2021 kern.notice kernel: [ 0.000000] random: get_random_bytes called from start_kernel+0x308/0x4fc with crng_init=0
Sat May 15 13:59:17 2021 kern.info kernel: [ 0.000000] CPU clock: 430.000 MHz
|
None: After build [[https://buildbot.openwrt.org/master/images/#/builders/1/builds/114|#114]] I tried to flash image r16730-bf8c16dfa2. Tried sysupgrade 1st from r16383-ec6293febc to r16730-bf8c16dfa2, which resulted in no boot. Then, I flashed it via TFTP with same result -> no boot. |
GKaURTbV: New patches for CVEs are already part of openwrt-21.02 branch - here is the FragAttacks commit to 21.02: Can you please try and see if a 21.02.0-SNAPSHOT works for you? Would be interesting, because like I said - I am running 21.02.0-rc1 on a TP-Link TL-WR1043ND v1 myself, but I need my device operational.. |
bartekm: Maybe I can shed some light here. Today I built the image for TP-Link WR1043ND v1 out of latest in 21.02 branch. Last weekend I built the image for the router out of master and still was able to observe the issue (System halted! at the boot time) however when I reverted this commit: 1b8bd17 and rebuilt and reflashed - the router booted successfully. |
None: Did the test as requested. Sysupgrade from master's r16383-ec6293febc to 21.02-SNAPSHOT r16086-f9b0215e28 was a success. System boots, everything I use functional. openwrt-21.02-snapshot-r16086-f9b0215e28-ath79-generic-tplink_tl-wr1043nd-v1-squashfs-sysupgrade.bin works like a charm. Please note, that I use the bare minimum. Even ipv6 disabled to free up some ram.
Mon May 17 00:04:34 2021 kern.notice kernel: [ 0.000000] Linux version 5.4.119 (builder@buildhost) (gcc version 8.4.0 (OpenWrt GCC 8.4.0 r16086-f9b0215e28)) #0 Sun May 16 22:02:14 2021
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] printk: bootconsole [early0] enabled
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] CPU0 revision is: 00019374 (MIPS 24Kc)
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] MIPS: machine is TP-Link TL-WR1043ND v1
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] SoC: Atheros AR9132 rev 2
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Initrd not found or empty - disabling initrd
Mon May 17 00:04:34 2021 kern.warn kernel: [ 0.000000] Primary instruction cache 64kB, VIPT, 4-way, linesize 32 bytes.
Mon May 17 00:04:34 2021 kern.warn kernel: [ 0.000000] Primary data cache 32kB, 4-way, VIPT, cache aliases, linesize 32 bytes
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Zone ranges:
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Normal [mem 0x0000000000000000-0x0000000001ffffff]
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Movable zone start for each node
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Early memory node ranges
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] node 0: [mem 0x0000000000000000-0x0000000001ffffff]
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Initmem setup node 0 [mem 0x0000000000000000-0x0000000001ffffff]
Mon May 17 00:04:34 2021 kern.debug kernel: [ 0.000000] On node 0 totalpages: 8192
Mon May 17 00:04:34 2021 kern.debug kernel: [ 0.000000] Normal zone: 72 pages used for memmap
Mon May 17 00:04:34 2021 kern.debug kernel: [ 0.000000] Normal zone: 0 pages reserved
Mon May 17 00:04:34 2021 kern.debug kernel: [ 0.000000] Normal zone: 8192 pages, LIFO batch:0
Mon May 17 00:04:34 2021 kern.debug kernel: [ 0.000000] pcpu-alloc: s0 r0 d32768 u32768 alloc=1*32768
Mon May 17 00:04:34 2021 kern.debug kernel: [ 0.000000] pcpu-alloc: [0] 0
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Built 1 zonelists, mobility grouping on. Total pages: 8120
Mon May 17 00:04:34 2021 kern.notice kernel: [ 0.000000] Kernel command line: console=ttyS0,115200 rootfstype=squashfs,jffs2
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Dentry cache hash table entries: 4096 (order: 2, 16384 bytes, linear)
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Inode-cache hash table entries: 2048 (order: 1, 8192 bytes, linear)
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Writing ErrCtl register=00000000
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Readback ErrCtl register=00000000
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] mem auto-init: stack:off, heap alloc:off, heap free:off
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] Memory: 24888K/32768K available (5222K kernel code, 192K rwdata, 684K rodata, 1188K init, 205K bss, 7880K reserved, 0K cma-reserved)
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] SLUB: HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] NR_IRQS: 51
Mon May 17 00:04:34 2021 kern.notice kernel: [ 0.000000] random: get_random_bytes called from 0x80657a28 with crng_init=0
Mon May 17 00:04:34 2021 kern.info kernel: [ 0.000000] CPU clock: 430.000 MHz
|
GKaURTbV: Fair enough. And thanks for thesting 21.02-SNAPSHOT! About the commit, which when reverted makes trunk work again - have you considered commenting about your findings directly to the author on github? And about the router itself - having both a TP-Link TL-WR1043ND v1 and an Archer C7 v5 in active use - to be honest, I am thinking of getting another C7 and letting the 1043 rest. |
GKaURTbV: One more thing - since you say you are building your own image, can you test applying the changes in pull request 4143 on top of trunk to see if that makes your 1043 work again? I think the above pull request fixes something missing in the commit you are reverting.. As a matter of fact, this missing pull request is the thing I am waiting on to try 21.02-SNAPSHOT myself. |
None: Okay, there must be a few misunderstandings here :)
|
mgondium: I confirm that reverting commit [[https://github.com/openwrt/openwrt/commit/1b8bd17c2d07c7397015a86941db9e114e1d349a]] makes trunk boot again on the 1043NDv1. I see that the problem is previous to commit [[https://github.com/openwrt/openwrt/commit/bf8c16dfa2cddd529f889510ca5fa5160606c325]], but Gergely Nemeth says build 114 didn't work... |
GKaURTbV: @PMGP can you apply patches on top of trunk and build/test? As I am understanding the issue - commit "ath79: lzma-loader: allow setting custom kernel magic" did almost everything right, except for setting some vars to empty, which if not done - get wrong values from previous device in tree or something like it. Commit 37b0a13 "ath79: set lzma-loader variables to null by default" from this pull request So if someone is building their own firmware - can you try applying the described changes and test? I use ready made firmwares too, so I cannot do it.. |
bartekm: I can confirm master starting from bf8c16d does not cause bricking wr-1043nd anymore. I built my own firmware a moment ago and successfully flash it using sysupgrade. |
GKaURTbV: Great news! But it makes me wonder why the "set to NULL" commit is only taken on the master branch, and why not take the CPE fix as well? Anyway, this means that trunk is NO LONGER bricking old ath9k devices (except CPE/WBS).. |
mgondium: I confirm that trunk r16765-6f3a05ebb0 is booting on the 1043NDv1. If this is due to the NULL variables commit I just don't understand why build 114 didn't work for Gergely Nemeth. Looks like this is SOLVED. |
None: Succesfully upgraded from 21.02-SNAPSHOT ( r16086-f9b0215e28 ) to master build [[https://buildbot.openwrt.org/master/images/#/builders/1/builds/127|#127]] ( r16794-d683175236 ). Anyway, I do confirm too, that master/trunk works again on 1043nd-v1.
Sun May 23 15:13:56 2021 kern.notice kernel: [ 0.000000] Linux version 5.4.121 (builder@buildhost) (gcc version 8.4.0 (OpenWrt GCC 8.4.0 r16794-d683175236)) #0 Sun M
ay 23 13:11:38 2021
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] printk: bootconsole [early0] enabled
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] CPU0 revision is: 00019374 (MIPS 24Kc)
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] MIPS: machine is TP-Link TL-WR1043ND v1
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] SoC: Atheros AR9132 rev 2
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Initrd not found or empty - disabling initrd
Sun May 23 15:13:56 2021 kern.warn kernel: [ 0.000000] Primary instruction cache 64kB, VIPT, 4-way, linesize 32 bytes.
Sun May 23 15:13:56 2021 kern.warn kernel: [ 0.000000] Primary data cache 32kB, 4-way, VIPT, cache aliases, linesize 32 bytes
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Zone ranges:
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Normal [mem 0x0000000000000000-0x0000000001ffffff]
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Movable zone start for each node
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Early memory node ranges
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] node 0: [mem 0x0000000000000000-0x0000000001ffffff]
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Initmem setup node 0 [mem 0x0000000000000000-0x0000000001ffffff]
Sun May 23 15:13:56 2021 kern.debug kernel: [ 0.000000] On node 0 totalpages: 8192
Sun May 23 15:13:56 2021 kern.debug kernel: [ 0.000000] Normal zone: 72 pages used for memmap
Sun May 23 15:13:56 2021 kern.debug kernel: [ 0.000000] Normal zone: 0 pages reserved
Sun May 23 15:13:56 2021 kern.debug kernel: [ 0.000000] Normal zone: 8192 pages, LIFO batch:0
Sun May 23 15:13:56 2021 kern.debug kernel: [ 0.000000] pcpu-alloc: s0 r0 d32768 u32768 alloc=1*32768
Sun May 23 15:13:56 2021 kern.debug kernel: [ 0.000000] pcpu-alloc: [0] 0
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Built 1 zonelists, mobility grouping on. Total pages: 8120
Sun May 23 15:13:56 2021 kern.notice kernel: [ 0.000000] Kernel command line: console=ttyS0,115200 rootfstype=squashfs,jffs2
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Dentry cache hash table entries: 4096 (order: 2, 16384 bytes, linear)
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Inode-cache hash table entries: 2048 (order: 1, 8192 bytes, linear)
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Writing ErrCtl register=00000000
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Readback ErrCtl register=00000000
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] mem auto-init: stack:off, heap alloc:off, heap free:off
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] Memory: 24312K/32768K available (5247K kernel code, 188K rwdata, 1192K rodata, 1236K init, 205K bss, 8456K res
erved, 0K cma-reserved)
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] SLUB: HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] NR_IRQS: 51
Sun May 23 15:13:56 2021 kern.notice kernel: [ 0.000000] random: get_random_bytes called from start_kernel+0x358/0x54c with crng_init=0
Sun May 23 15:13:56 2021 kern.info kernel: [ 0.000000] CPU clock: 430.000 MHz
|
mgondium:
Trunk is bricking old ath9k devices, so far a TPLINK MR3220v1 and a WR1043NDv1, retested with a fresh buildroot and base configuration, same outcome.
Autobooting in 1 seconds## Booting image at bf020000 ... Uncompressing Kernel Image ... OK
Starting kernel ...
OpenWrt kernel loader for AR7XXX/AR9XXX
Copyright (C) 2011 Gabor Juhos juhosg@openwrt.org
Looking for OpenWrt image... not found!
System halted!
The text was updated successfully, but these errors were encountered: