On 2 August 2018 at 19:52, Matt Hart matthew.hart@linaro.org wrote:
On 2 August 2018 at 18:46, Ard Biesheuvel ard.biesheuvel@linaro.org wrote:
On 2 August 2018 at 19:14, Mark Brown broonie@kernel.org wrote:
On Thu, Aug 02, 2018 at 09:44:10AM -0700, kernelci.org bot wrote:
Today's -next fails to boot on SynQuacer when booted with ACPI:
arm64: defconfig: synquacer-acpi: lab-mhart: new failure (last pass: next-20180801)
The boot seems to hang after enumerating the disks:
13:17:59.325408 [ 7.416795] ata1: SATA link up 6.0 Gbps (SStatus 133 SControl 300) 13:17:59.331425 [ 7.423625] ata1.00: ATA-10: WDC WD10EZEX-22MFCA0, 01.01A01, max UDMA/133 13:17:59.337401 [ 7.430415] ata1.00: 1953525168 sectors, multi 16: LBA48 NCQ (depth 32), AA 13:17:59.343411 [ 7.438058] ata1.00: configured for UDMA/133 13:17:59.349409 [ 7.442752] scsi 0:0:0:0: Direct-Access ATA WDC WD10EZEX-22M 1A01 PQ: 0 ANSI: 5 13:17:59.360397 [ 7.451387] sd 0:0:0:0: [sda] 1953525168 512-byte logical blocks: (1.00 TB/932 GiB) 13:17:59.366427 [ 7.459059] sd 0:0:0:0: [sda] 4096-byte physical blocks 13:17:59.367152 [ 7.464324] sd 0:0:0:0: [sda] Write Protect is off 13:17:59.377343 [ 7.469161] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA 13:17:59.431420 [ 7.526763] sda: sda1 sda2 sda3 13:17:59.435383 [ 7.531608] sd 0:0:0:0: [sda] Attached SCSI disk 13:17:59.671358 [ 7.763087] ata2: SATA link down (SStatus 0 SControl 300)
In a working boot like yesterday's we then see networking output which seems to be missing here:
10:32:34.187752 [ 8.362173] netsec SCX0001:00 eth0: Link is Down 10:32:38.325575 [ 12.463251] netsec SCX0001:00 eth0: Link is Up - 1Gbps/Full - flow control rx/tx 10:32:38.345614 [ 12.485944] IP-Config: Complete:
All the info we have including full logs is here:
https://kernelci.org/boot/id/5b63052b59b5143ca896bab8/
and I kicked off a bisect:
https://ci.kernelci.org/job/lava-bisect/5624/
which will take a while to run, I'll report back tomorrow.
(you can only see if you have an account, sorry - we can probably sort you out with a login if you like, or people on #kernelci should be able to at least look at the results for you.)
I had a quick look at the delta between yesterday's and today's next, and the only thing that really stands out are the anon VMA changes in mainline.
It would be useful to compare with synquacer-dtb, though. Is that still online?
It's been hard to keep that Synquacer machine online as the serial just quietly drops out. It's still connected over USB but no output comes from the board. If I pull the usb serial and reconnect it, it all works fine for a few days until it stops again. I'm planning to put an RS232 convertor in the back to give a normal serial port and see if that improves it.
On the rev 0.1 boards, this was all working fine, but the part was EOL so they had to make a last minute change. On my rev 0.2 board, I don't use the USB serial at all because they screwed something up and I can't leave it bus powered if I power down the board. It sounds like they didn't quite get it right on rev 0.3 either.
You can actually use a PCI serial adapter, and it will work from UEFI onwards (after you enable it for console output from the UEFI menu). It does require a console=ttyS0 kernel command line argument though