linux
linux copied to clipboard
raspberrypi
usb/dwc2: Set correct state on gadget disconnect
When operating as a USB device, when the device is disconnected the suspend interrupt is called rather than the disconnect interrupt, this results in the state remaining as "configured". This change changes the state to "not attached" instead when the suspend interrupt is called.
I don't know whether this patch interferes with anything else or whether it's the correct way of "fixing" this issue, there is chatter on the web about similar behaviour a long time ago but patches were submitted, so I don't know whether this is something which is just affecting the hardware implementation on the Pi.
Without the correct state being reflected, it's not possible to detect when a gadget is connected and disconnected from a device as without this patch after enumeration even pulling the the cable the Pi device remains in "configured" state. With this patch it reverts back to "not attached".
Please provide some more information about the issue:
- Which Raspberry Pi model did you use?
- How do you power the Raspberry Pi?
- Did you tried a recent mainline kernel?
Pi Zero W, powered via a pi power supply. Only tried the most recent kernel available through raspbian. This issue has been mentioned on the Pi forums a few times in the past without any resolution, also searching the web throws up results.
Let me rephrase my second question, which interface powers the Raspberry Pi Zero W: a) GPIO header b) USB c) PWR IN
dwc2 is under active development in the mainline kernel. So your patch should be submitted to linux-usb for a proper solution. But you're patch must be tested with a recent mainline kernel like Linux 5.2 before.
PWR IN.
I’m unsure whether this issue affects anything other than raspberry pi’s (external hardware around the OTG section, my colleague made a slight “improvement” on the OTG section on one of our boards which had unintended consequences with host/device detection so I know its always a possibility), it would seem somewhat of an oversight if it affected everything and I’d expect more results when searching for the issue, there are some generic search hits for it from a couple of years back but patches were submitted.
I’ll see if I can try with a mainline kernel, I. don't suppose you have any links on setting up a system using such a beast? I suppose i could also try pulling the latest source for the dwc2 module and see if it compiles with my kernel.
I only posted this “fix” as i posted a solution to this problem on the forums and one of the staff suggested i create a pull request so people can look at it.
Regardless of my solution, its a problem on the Pi Zero W (and potentially other Pi’s) which means that you cant track the disconnected and enumerated state. I’ll admit it was a quick hack fix, but it works for me in my situation, at the very least hopefully it can cause some discussion about the problem.
Please try to use this gist in order to compile the mainline kernel. You can use bcm2835_defconfig as config and bcm2835-rpi-zero-w.dtb as devicetree binary. Do not use the devicetree from Raspbian in combination with a mainline kernel.
ok, will take a look and have a go.
For what its worth, I’ve just looked at the history for the dwc2 module and can’t see any commits that look like they would have any effect on this, but I guess you never know until you try, ot holding my breath.
@lategoodbye Hi, I've followed the instructions with the latest mainline, kernel modules installed, device tree overlay copied and selected.
Kernel starts to boot and then stops at "Waiting for root device"
Switch back to original kernel, system boots fine.
Any ideas?
I also tried getting the .confid out of the booted rapbian system and used that as the config for the kernel when building, still have same issue.
I also tried getting the .confid out of the booted rapbian system and used that as the config for the kernel when building, still have same issue.
This won't work, you will need to replace kernel and DTB. You will need to use make bcm2835_defconfig as config. Overlays doesn't work yet in mainline.
Maybe you can provide your dmesg and config.txt.
ok, will try again.
No boot log, I've just ordered a USB to TTL adaptor so I can hook up a terminal to capture everything.
I'll repull the repo and build from fresh and try again.
Thanks for your support, much appreciated!
Btw are you using the stable tree e.g. git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git ?
No, I was using the version that was in the gist, I'm just pulling the stable now.
The torvalds tree is also okay, you will need to choose a final release like v5.2.0
Ok, so I have kernel booted, but a lot of stuff appears not to work, I don't seem to be able to load the dwc2 module and there's a whole load of errors about missing firmware.
[ 0.000000] Booting Linux on physical CPU 0x0
[ 0.000000] Linux version 5.3.0-rc4 (adrian@adrian-Standard-PC-Q35-ICH9-2009) (gcc version 7.3.1 20180425 [linaro-7.3-2018.05 revision d29120a424ecfbc167ef90065c0eeb7f91977701] (Linaro GCC 7.3-2018.05)) #1 Thu Aug 15 12:43:40 BST 2019
[ 0.000000] CPU: ARMv6-compatible processor [410fb767] revision 7 (ARMv7), cr=00c5387d
[ 0.000000] CPU: PIPT / VIPT nonaliasing data cache, VIPT nonaliasing instruction cache
[ 0.000000] OF: fdt: Machine model: Raspberry Pi Zero W Rev 1.1
[ 0.000000] Memory policy: Data cache writeback
[ 0.000000] cma: Reserved 32 MiB at 0x19c00000
[ 0.000000] On node 0 totalpages: 114688
[ 0.000000] Normal zone: 896 pages used for memmap
[ 0.000000] Normal zone: 0 pages reserved
[ 0.000000] Normal zone: 114688 pages, LIFO batch:31
[ 0.000000] CPU: All CPU(s) started in SVC mode.
[ 0.000000] pcpu-alloc: s0 r0 d32768 u32768 alloc=1*32768
[ 0.000000] pcpu-alloc: [0] 0
[ 0.000000] Built 1 zonelists, mobility grouping on. Total pages: 113792
[ 0.000000] Kernel command line: video=HDMI-A-1:1920x1080@60 dma.dmachans=0x7f35 bcm2708.boardrev=0x9000c1 bcm2708.serial=0x7a1f2d6c bcm2708.uart_clock=48000000 bcm2708.disk_led_gpio=47 smsc95xx.macaddr=B8:27:EB:1F:2D:6C vc_mem.mem_base=0x1ec00000 vc_mem.mem_size=0x20000000 dwc_otg.lpm_enable=0 console=ttyS0,115200 console=tty1 root=PARTUUID=99c633c0-02 rootfstype=ext4 rootwait elevator=deadline fsck.repair=yes debug
[ 0.000000] Dentry cache hash table entries: 65536 (order: 6, 262144 bytes, linear)
[ 0.000000] Inode-cache hash table entries: 32768 (order: 5, 131072 bytes, linear)
[ 0.000000] mem auto-init: stack:off, heap alloc:off, heap free:off
[ 0.000000] Memory: 408244K/458752K available (8192K kernel code, 611K rwdata, 2644K rodata, 1024K init, 679K bss, 17740K reserved, 32768K cma-reserved)
[ 0.000000] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
[ 0.000000] ftrace: allocating 28438 entries in 56 pages
[ 0.000000] NR_IRQS: 16, nr_irqs: 16, preallocated irqs: 16
[ 0.000000] random: get_random_bytes called from start_kernel+0x2dc/0x4f0 with crng_init=0
[ 0.000024] sched_clock: 32 bits at 1000kHz, resolution 1000ns, wraps every 2147483647500ns
[ 0.000059] clocksource: timer: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 1911260446275 ns
[ 0.000152] bcm2835: system timer (irq = 27)
[ 0.000833] Console: colour dummy device 80x30
[ 0.001567] printk: console [tty1] enabled
[ 0.001661] Calibrating delay loop... 697.95 BogoMIPS (lpj=3489792)
[ 0.060331] pid_max: default: 32768 minimum: 301
[ 0.060846] Mount-cache hash table entries: 1024 (order: 0, 4096 bytes, linear)
[ 0.060913] Mountpoint-cache hash table entries: 1024 (order: 0, 4096 bytes, linear)
[ 0.062254] CPU: Testing write buffer coherency: ok
[ 0.063567] Setting up static identity map for 0x100000 - 0x100054
[ 0.069014] devtmpfs: initialized
[ 0.078267] VFP support v0.3: implementor 41 architecture 1 part 20 variant b rev 5
[ 0.078752] clocksource: jiffies: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 19112604462750000 ns
[ 0.078833] futex hash table entries: 256 (order: -1, 3072 bytes, linear)
[ 0.083552] pinctrl core: initialized pinctrl subsystem
[ 0.084529] NET: Registered protocol family 16
[ 0.087195] DMA: preallocated 256 KiB pool for atomic coherent allocations
[ 0.092892] No ATAGs?
[ 0.092923] hw-breakpoint: found 6 breakpoint and 1 watchpoint registers.
[ 0.093011] hw-breakpoint: maximum watchpoint size is 4 bytes.
[ 0.093365] Serial: AMBA PL011 UART driver
[ 0.132174] SCSI subsystem initialized
[ 0.132717] usbcore: registered new interface driver usbfs
[ 0.132882] usbcore: registered new interface driver hub
[ 0.133091] usbcore: registered new device driver usb
[ 0.133751] videodev: Linux video capture interface: v2.00
[ 0.134170] Advanced Linux Sound Architecture Driver Initialized.
[ 0.135084] Bluetooth: Core ver 2.22
[ 0.135229] NET: Registered protocol family 31
[ 0.135267] Bluetooth: HCI device and connection manager initialized
[ 0.135324] Bluetooth: HCI socket layer initialized
[ 0.135367] Bluetooth: L2CAP socket layer initialized
[ 0.135435] Bluetooth: SCO socket layer initialized
[ 0.136794] clocksource: Switched to clocksource timer
[ 0.366786] simple-framebuffer 1e887000.framebuffer: framebuffer at 0x1e887000, 0x373800 bytes, mapped to 0x(ptrval)
[ 0.366996] simple-framebuffer 1e887000.framebuffer: format=r5g6b5, mode=1824x984x16, linelength=3648
[ 0.398738] Console: switching to colour frame buffer device 228x61
[ 0.429263] simple-framebuffer 1e887000.framebuffer: fb0: simplefb registered!
[ 0.449029] thermal_sys: Registered thermal governor 'step_wise'
[ 0.449767] NET: Registered protocol family 2
[ 0.451264] tcp_listen_portaddr_hash hash table entries: 512 (order: 0, 4096 bytes, linear)
[ 0.451602] TCP established hash table entries: 4096 (order: 2, 16384 bytes, linear)
[ 0.451934] TCP bind hash table entries: 4096 (order: 2, 16384 bytes, linear)
[ 0.452236] TCP: Hash tables configured (established 4096 bind 4096)
[ 0.452627] UDP hash table entries: 256 (order: 0, 4096 bytes, linear)
[ 0.452879] UDP-Lite hash table entries: 256 (order: 0, 4096 bytes, linear)
[ 0.453411] NET: Registered protocol family 1
[ 0.454677] RPC: Registered named UNIX socket transport module.
[ 0.454911] RPC: Registered udp transport module.
[ 0.455076] RPC: Registered tcp transport module.
[ 0.455237] RPC: Registered tcp NFSv4.1 backchannel transport module.
[ 0.458195] hw perfevents: no irqs for PMU, sampling events not supported
[ 0.458523] hw perfevents: enabled with armv6_1176 PMU driver, 3 counters available
[ 0.461356] Initialise system trusted keyrings
[ 0.461983] workingset: timestamp_bits=30 max_order=17 bucket_order=0
[ 0.478550] Installing knfsd (copyright (C) 1996 [email protected]).
[ 0.541206] jitterentropy: Initialization failed with host not compliant with requirements: 2
[ 0.541629] Key type asymmetric registered
[ 0.541794] Asymmetric key parser 'x509' registered
[ 0.542086] Block layer SCSI generic (bsg) driver version 0.4 loaded (major 252)
[ 0.542337] io scheduler mq-deadline registered
[ 0.542500] io scheduler kyber registered
[ 0.552466] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled
[ 0.556369] 20215040.serial: ttyS1 at MMIO 0x0 (irq = 53, base_baud = 31250000) is a 16550
[ 0.562707] bcm2835-power bcm2835-power: Broadcom BCM2835 power domains driver
[ 0.564330] libphy: Fixed MDIO Bus: probed
[ 0.565041] usbcore: registered new interface driver zd1211rw
[ 0.565358] usbcore: registered new interface driver lan78xx
[ 0.575057] usbcore: registered new interface driver asix
[ 0.584568] usbcore: registered new interface driver ax88179_178a
[ 0.593811] usbcore: registered new interface driver cdc_ether
[ 0.603087] usbcore: registered new interface driver smsc95xx
[ 0.612069] usbcore: registered new interface driver net1080
[ 0.620950] usbcore: registered new interface driver cdc_subset
[ 0.629688] usbcore: registered new interface driver zaurus
[ 0.638462] usbcore: registered new interface driver cdc_ncm
[ 0.647837] usbcore: registered new interface driver usb-storage
[ 0.656577] i2c /dev entries driver
[ 0.668135] i2c-bcm2835 20805000.i2c: Could not read clock-frequency property
[ 0.679307] bcm2835-wdt bcm2835-wdt: Broadcom BCM2835 watchdog timer
[ 0.688973] sdhci: Secure Digital Host Controller Interface driver
[ 0.697554] sdhci: Copyright(c) Pierre Ossman
[ 0.802653] sdhost-bcm2835 20202000.mmc: loaded - DMA enabled (>1)
[ 0.811789] sdhci-pltfm: SDHCI platform and OF driver helper
[ 0.821434] sdhci-iproc 20300000.sdhci: allocated mmc-pwrseq
[ 0.862499] mmc1: SDHCI controller on 20300000.sdhci [20300000.sdhci] using PIO
[ 0.873453] ledtrig-cpu: registered to indicate activity on CPUs
[ 0.883463] usbcore: registered new interface driver usbhid
[ 0.892348] usbhid: USB HID core driver
[ 0.902675] bcm2835-mbox 2000b880.mailbox: mailbox enabled
[ 0.915830] oprofile: using arm/armv6
[ 0.931413] NET: Registered protocol family 10
[ 0.942069] mmc1: queuing unknown CIS tuple 0x80 (2 bytes)
[ 0.952598] Segment Routing with IPv6
[ 0.962213] sit: IPv6, IPv4 and MPLS over IPv4 tunneling driver
[ 0.971997] mmc1: queuing unknown CIS tuple 0x80 (3 bytes)
[ 0.982025] NET: Registered protocol family 17
[ 0.993327] mmc1: queuing unknown CIS tuple 0x80 (3 bytes)
[ 1.002442] random: fast init done
[ 1.012517] Loading compiled-in X.509 certificates
[ 1.023749] mmc1: queuing unknown CIS tuple 0x80 (7 bytes)
[ 1.039305] mmc0: host does not support reading read-only switch, assuming write-enable
[ 1.051904] 20201000.serial: ttyAMA0 at MMIO 0x20201000 (irq = 81, base_baud = 0) is a PL011 rev2
[ 1.061148] mmc0: new high speed SDHC card at address aaaa
[ 1.070453] serial serial0: tty port ttyAMA0 registered
[ 1.081625] mmcblk0: mmc0:aaaa SC16G 14.8 GiB
[ 1.093850] raspberrypi-firmware soc:firmware: Attached to firmware from 2019-07-09 14:40
[ 1.108752] mmcblk0: p1 p2
[ 1.119977] debugfs: Directory '20902000.hdmi' with parent 'vc4-hdmi' already present!
[ 1.132270] vc4_hdmi 20902000.hdmi: ASoC: Failed to create component debugfs directory: -17
[ 1.143143] vc4_hdmi 20902000.hdmi: vc4-hdmi-hifi <-> 20902000.hdmi mapping ok
[ 1.154190] vc4-drm soc:gpu: bound 20902000.hdmi (ops vc4_hdmi_ops)
[ 1.163622] vc4-drm soc:gpu: bound 20806000.vec (ops vc4_vec_ops)
[ 1.172915] vc4-drm soc:gpu: bound 20004000.txp (ops vc4_txp_ops)
[ 1.181970] vc4-drm soc:gpu: bound 20400000.hvs (ops vc4_hvs_ops)
[ 1.191974] vc4-drm soc:gpu: bound 20206000.pixelvalve (ops vc4_crtc_ops)
[ 1.202089] vc4-drm soc:gpu: bound 20207000.pixelvalve (ops vc4_crtc_ops)
[ 1.211996] vc4-drm soc:gpu: bound 20807000.pixelvalve (ops vc4_crtc_ops)
[ 1.220724] vc4-drm soc:gpu: bound 20c00000.v3d (ops vc4_v3d_ops)
[ 1.229129] checking generic (1e887000 373800) vs hw (0 ffffffff)
[ 1.237367] fb0: switching to vc4drmfb from simple
[ 1.246124] Console: switching to colour dummy device 80x30
[ 1.246354] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 1.246395] [drm] Driver supports precise vblank timestamp query.
[ 1.248797] [drm] Initialized vc4 0.0.0 20140616 for soc:gpu on minor 0
[ 1.338907] Console: switching to colour frame buffer device 240x67
[ 1.389373] vc4-drm soc:gpu: fb0: vc4drmfb frame buffer device
[ 1.414017] mmc1: new high speed SDIO card at address 0001
[ 1.637039] dwc2 20980000.usb: EPs: 8, dedicated fifos, 4080 entries in SPRAM
[ 1.638886] dwc2 20980000.usb: DWC OTG Controller
[ 1.639173] dwc2 20980000.usb: new USB bus registered, assigned bus number 1
[ 1.639508] dwc2 20980000.usb: irq 33, io mem 0x20980000
[ 1.641555] hub 1-0:1.0: USB hub found
[ 1.641883] hub 1-0:1.0: 1 port detected
[ 1.646472] cfg80211: Loading compiled-in X.509 certificates for regulatory database
[ 1.654180] cfg80211: Loaded X.509 cert 'sforshee: 00b28ddf47aef9cea7'
[ 1.654548] ALSA device list:
[ 1.654951] platform regulatory.0: Direct firmware load for regulatory.db failed with error -2
[ 1.655259] cfg80211: failed to load regulatory.db
[ 1.655676] #0: vc4-hdmi
[ 1.673072] EXT4-fs (mmcblk0p2): mounted filesystem with ordered data mode. Opts: (null)
[ 1.673478] VFS: Mounted root (ext4 filesystem) readonly on device 179:2.
[ 1.682967] devtmpfs: mounted
[ 1.689573] Freeing unused kernel memory: 1024K
[ 1.690207] Run /sbin/init as init process
[ 2.479109] systemd[1]: System time before build time, advancing clock.
[ 2.531218] systemd[1]: Failed to find module 'autofs4'
[ 2.578948] systemd[1]: systemd 241 running in system mode. (+PAM +AUDIT +SELINUX +IMA +APPARMOR +SMACK +SYSVINIT +UTMP +LIBCRYPTSETUP +GCRYPT +GNUTLS +ACL +XZ +LZ4 +SECCOMP +BLKID +ELFUTILS +KMOD -IDN2 +IDN -PCRE2 default-hierarchy=hybrid)
[ 2.580022] systemd[1]: No virtualization found in DMI
[ 2.580325] systemd[1]: No virtualization found in CPUID
[ 2.580699] systemd[1]: Virtualization XEN not found, /proc/xen does not exist
[ 2.581687] systemd[1]: No virtualization found in /proc/device-tree/*
[ 2.582539] systemd[1]: UML virtualization not found in /proc/cpuinfo.
[ 2.582947] systemd[1]: This platform does not support /proc/sysinfo
[ 2.583271] systemd[1]: Found VM virtualization none
[ 2.583559] systemd[1]: Detected architecture arm.
[ 2.585159] systemd[1]: Mounting cgroup to /sys/fs/cgroup/perf_event of type cgroup with options perf_event.
[ 4.106401] systemd-bless-boot-generator[81]: Skipping generator, not an EFI boot.
[ 4.235710] systemd-fstab-generator[84]: Parsing /etc/fstab
[ 4.256683] systemd-rc-local-generator[88]: Automatically adding rc-local.service.
[ 4.301950] systemd-hibernate-resume-generator[87]: Not running in an initrd, quitting.
[ 4.329244] systemd-gpt-auto-generator[86]: Failed to chase block device '/', ignoring: No such file or directory
[ 4.393309] systemd-fstab-generator[84]: Found entry what=proc where=/proc type=proc makefs=no nofail=no noauto=no
[ 4.412827] systemd-rc-local-generator[88]: /usr/sbin/halt.local does not exist, skipping.
[ 4.429612] systemd-sysv-generator[91]: Native unit for nfs-common.service already exists, skipping.
[ 4.440437] systemd-gpt-auto-generator[86]: mmcblk0p2: Root device /dev/mmcblk0.
[ 4.451929] systemd-fstab-generator[84]: Found entry what=/dev/disk/by-partuuid/99c633c0-01 where=/boot type=vfat makefs=no nofail=no noauto=no
[ 4.472301] systemd-sysv-generator[91]: Native unit for rpcbind.service already exists, skipping.
[ 4.490248] systemd-sysv-generator[91]: Native unit for hwclock.service already exists, skipping.
[ 4.524009] systemd-sysv-generator[91]: Native unit for fake-hwclock.service already exists, skipping.
[ 4.541920] systemd-fstab-generator[84]: Found entry what=/dev/disk/by-partuuid/99c633c0-02 where=/ type=ext4 makefs=no nofail=no noauto=no
[ 4.563265] systemd-sysv-generator[91]: Native unit for networking.service already exists, skipping.
[ 4.591848] systemd-sysv-generator[91]: Native unit for bluetooth.service already exists, skipping.
[ 4.613160] systemd-sysv-generator[91]: Native unit for paxctld.service already exists, skipping.
[ 4.635653] systemd-sysv-generator[91]: Native unit for cron.service already exists, skipping.
[ 4.654136] systemd-sysv-generator[91]: Native unit for dphys-swapfile.service already exists, skipping.
[ 4.672632] systemd-sysv-generator[91]: Native unit for dbus.service already exists, skipping.
[ 4.695208] systemd-gpt-auto-generator[86]: No suitable partition table found, ignoring.
[ 4.799236] printk: systemd-sysv-ge: 69 output lines suppressed due to ratelimiting
[ 5.837606] random: systemd: uninitialized urandom read (16 bytes read)
[ 5.902079] random: systemd: uninitialized urandom read (16 bytes read)
[ 5.931773] random: systemd: uninitialized urandom read (16 bytes read)
[ 7.277886] systemd[114]: Operating on architecture: arm
[ 7.283699] systemd[114]: Operating on architecture: arm
[ 7.284765] systemd[114]: Operating on architecture: arm
[ 7.285722] systemd[114]: Restricting namespace to: .
[ 7.285961] systemd[114]: Operating on architecture: arm
[ 7.286248] systemd[114]: Blocking cgroup.
[ 7.286519] systemd[114]: Blocking ipc.
[ 7.306936] systemd[114]: Blocking net.
[ 7.307290] systemd[114]: Blocking mnt.
[ 7.307588] systemd[114]: Blocking pid.
[ 7.307909] systemd[114]: Blocking user.
[ 7.308227] systemd[114]: Blocking uts.
[ 7.309969] systemd[114]: Operating on architecture: arm
[ 7.838758] systemd[1]: libmount event [rescan: yes]
[ 7.875576] systemd[1]: run-rpc_pipefs.mount: Changed mounting -> mounting-done
[ 7.927309] systemd[1]: sys-kernel-debug.mount: Changed mounting -> mounting-done
[ 7.957678] systemd[1]: Received SIGCHLD from PID 96 (mount).
[ 7.986625] systemd[1]: Child 96 (mount) died (code=exited, status=0/SUCCESS)
[ 8.012369] systemd-journald[114]: Found cgroup2 on /sys/fs/cgroup/unified, unified hierarchy for systemd controller
[ 8.022756] systemd[1]: run-rpc_pipefs.mount: Child 96 belongs to run-rpc_pipefs.mount.
[ 8.037122] systemd[1]: run-rpc_pipefs.mount: Mount process exited, code=exited, status=0/SUCCESS
[ 8.046687] systemd[1]: run-rpc_pipefs.mount: Changed mounting-done -> mounted
[ 8.088060] systemd-journald[114]: Journal effective settings seal=no compress=yes compress_threshold_bytes=512B
[ 8.097923] systemd[1]: run-rpc_pipefs.mount: Job 70 run-rpc_pipefs.mount/start finished, result=done
[ 8.118088] systemd-journald[114]: Fixed min_use=1.0M max_use=21.5M max_size=2.6M min_size=512.0K keep_free=32.3M n_max_files=100
[ 8.128013] systemd[1]: Mounted RPC Pipe File System.
[ 8.151677] systemd-journald[114]: Reserving 4913 entries in hash table.
[ 8.188261] systemd-journald[114]: Vacuuming...
[ 8.217564] systemd-journald[114]: Vacuuming done, freed 0B of archived journals from /run/log/journal/88bb7e0266774e40a3986e7d17fc46ff.
[ 8.217839] systemd-journald[114]: Flushing /dev/kmsg...
[ 9.397082] systemd-journald[114]: systemd-journald running as pid 114
[ 9.462267] systemd-journald[114]: Sent READY=1 notification.
[ 9.462657] systemd-journald[114]: Sent WATCHDOG=1 notification.
[ 11.083266] EXT4-fs (mmcblk0p2): re-mounted. Opts: (null)
[ 15.050776] vchiq: module is from the staging directory, the quality is unknown, you have been warned.
[ 15.167155] vchiq: vchiq_init_state: slot_zero = (ptrval)
[ 15.380466] bcm2835-rng 20104000.rng: hwrng registered
[ 15.884864] Bluetooth: HCI UART driver ver 2.3
[ 15.884887] Bluetooth: HCI UART protocol H4 registered
[ 15.897253] uart-pl011 20201000.serial: no DMA platform data
[ 16.048302] Bluetooth: hci0: BCM: chip id 94
[ 16.048988] Bluetooth: hci0: BCM: features 0x2e
[ 16.050897] Bluetooth: hci0: BCM43430A1
[ 16.050936] Bluetooth: hci0: BCM43430A1 (001.002.009) build 0000
[ 16.051885] Bluetooth: HCI UART protocol Broadcom registered
[ 16.807651] Bluetooth: hci0: BCM43430A1 (001.002.009) build 0360
[ 17.967345] brcmfmac: brcmf_fw_alloc_request: using brcm/brcmfmac43430-sdio for chip BCM43430/1
[ 17.990124] brcmfmac mmc1:0001:1: Direct firmware load for brcm/brcmfmac43430-sdio.raspberrypi,model-zero-w.txt failed with error -2
[ 18.270284] brcmfmac: brcmf_fw_alloc_request: using brcm/brcmfmac43430-sdio for chip BCM43430/1
[ 18.270526] brcmfmac: brcmf_c_process_clm_blob: no clm_blob available (err=-2), device may have limited channels available
[ 18.271942] brcmfmac: brcmf_c_preinit_dcmds: Firmware: BCM43430/1 wl0: Oct 23 2017 03:55:53 version 7.45.98.38 (r674442 CY) FWID 01-e58d219f
[ 23.403367] snd_bcm2835: module is from the staging directory, the quality is unknown, you have been warned.
[ 23.956250] bcm2835_audio bcm2835_audio: card created with 8 channels
[ 24.856324] bcm2835_v4l2: module is from the staging directory, the quality is unknown, you have been warned.
[ 25.287132] systemd[1]: Pulling in systemd-fsck-root.service/start from local-fs.target/start
[ 25.337080] systemd[1]: Pulling in shutdown.target/stop from local-fs.target/start
[ 25.357046] systemd[1]: Pulling in emergency.target/stop from sysinit.target/start
[ 25.377094] systemd[1]: Pulling in dbus.socket/start from bluetooth.service/start
[ 25.427102] systemd[1]: Pulling in -.mount/start from dbus.socket/start
[ 25.467225] systemd[1]: Pulling in shutdown.target/stop from bluetooth.target/start
[ 25.537106] systemd[1]: Found redundant job systemd-journal-flush.service/start, dropping from transaction.
[ 25.557036] systemd[1]: Found redundant job systemd-udevd.service/start, dropping from transaction.
[ 25.577107] systemd[1]: Found redundant job boot.mount/start, dropping from transaction.
[ 25.667326] systemd[1]: Found redundant job -.slice/start, dropping from transaction.
[ 29.567005] random: crng init done
[ 29.567031] random: 7 urandom warning(s) missed due to ratelimiting
[ 33.858720] systemd[1]: Accepted new private connection.
[ 33.882561] systemd[1]: systemd-journald.service: Got notification message from PID 114 (FDSTORE=1)
[ 33.897079] systemd[1]: systemd-journald.service: Added fd 58 (n/a) to fd store.
[ 33.927146] systemd[1]: systemd-journald.service: Received EPOLLHUP on stored fd 46 (stored), closing.
[ 33.957050] systemd[1]: systemd-journald.service: Received EPOLLHUP on stored fd 50 (stored), closing.
[ 33.977056] systemd[1]: Sent message type=method_call sender=n/a destination=org.freedesktop.DBus path=/org/freedesktop/DBus interface=org.freedesktop.DBus member=AddMatch cookie=2 reply_cookie=0 signature=s error-name=n/a error-message=n/a
[ 33.997089] systemd[1]: Sent message type=method_call sender=n/a destination=org.freedesktop.DBus path=/org/freedesktop/DBus interface=org.freedesktop.DBus member=AddMatch cookie=3 reply_cookie=0 signature=s error-name=n/a error-message=n/a
[ 34.017102] systemd[1]: Sent message type=method_call sender=n/a destination=org.freedesktop.DBus path=/org/freedesktop/DBus interface=org.freedesktop.DBus member=AddMatch cookie=4 reply_cookie=0 signature=s error-name=n/a error-message=n/a
[ 34.037106] systemd[1]: Sent message type=method_call sender=n/a destination=org.freedesktop.DBus path=/org/freedesktop/DBus interface=org.freedesktop.DBus member=AddMatch cookie=6 reply_cookie=0 signature=s error-name=n/a error-message=n/a
[ 34.057041] systemd[1]: Sent message type=method_call sender=n/a destination=org.freedesktop.DBus path=/org/freedesktop/DBus interface=org.freedesktop.DBus member=RequestName cookie=7 reply_cookie=0 signature=su error-name=n/a error-message=n/a
[ 34.608059] Adding 102396k swap on /var/swap. Priority:-2 extents:1 across:102396k SS
[ 36.483529] IPv6: ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
[ 80.357032] systemd[1]: Sent message type=signal sender=n/a destination=n/a path=/org/freedesktop/systemd1 interface=org.freedesktop.systemd1.Manager member=UnitNew cookie=202 reply_cookie=0 signature=so error-name=n/a error-message=n/a
Congratulation, you successful compiled a mainline kernel. dwc2 is working fine (no module, it's compiled in). Don't care about Wifi firmware, it's not relevant for the issue.
Are you able to reproduce the issue or are the sysfs entries missing?
ok, f_hid is disabled in this kernel, time to build another kernel image.
Everything else looks good, UDC is appearing.
@lategoodbye Same issue, disconnect the cable and the Pi Zero W still thinks it's in a configured state.
pi@raspberrypi:/sys/class/udc/20980000.usb $ cat state
configured
<cable pulled>
<1 minute later>
pi@raspberrypi:/sys/class/udc/20980000.usb $ cat state
configured
It's definitely broken.
"Great". Please apply your patch to your mainline repo and retest.
After that you need to report this issue (including reference to this issue and your inlined patch) via email to the following addresses: Minas Harutyunyan [email protected] Felipe Balbi [email protected] Greg Kroah-Hartman [email protected] Stefan Wahren [email protected] [email protected]
Thanks
Ok, let me do a little more testing and reading and if I need to amend I will, then I'll submit a patch.
What happens if this issue only occurs on the Pi? I guess the people looking at the patch will try without it first to see they can reproduce it on their hardware?
It's a bug and you were able to reproduce with a mainline kernel reliable. Don't be afraid this might be specific to the Pi (which i don't believe). It's job of the maintainers to decide, how to fix this. So don't waste too much time with preparing the "perfect" patch. This can be done afterwards.
It's a bug and you were able to reproduce with a mainline kernel reliable. Don't be afraid this might be specific to the Pi (which i don't believe). It's job of the maintainers to decide, how to fix this. So don't waste too much time with preparing the "perfect" patch. This can be done afterwards.
Ok, thanks for all your time, help and advice, it's been much appreciated, I've never really delved into the kernel on linux (I did on the now dead ucLinux where I wrote some drivers for stuff).
This particular bug is particularly annoying in that I found it straight away while conducting tests for using a Pi Zero W as a USB gadget (It's actually a bridge to make a Bluetooth Keyboard into a USB HID keyboard so I an use it with ESXi) and while I stormed through all the initial tests, enumeration, getting the actual "bridge" working I then hit a roadblock when it came to detecting whether the USB device was connected, bluetooth I get udev events, but USB requires monitoring that state file and that's when the trouble started.
It's a bug and you were able to reproduce with a mainline kernel reliable. Don't be afraid this might be specific to the Pi (which i don't believe). It's job of the maintainers to decide, how to fix this. So don't waste too much time with preparing the "perfect" patch. This can be done afterwards.
Quick question, I hooked up a serial console and while it works with the raspbian kernel I get no output when booting my 5.3 kernel, is this a limitation of the mainline kernel?
I looked about the config and everything that should be enabled (AFAIK) is enabled.
Ahh, it moves the serial port. Changed by console parameter to console=ttyS1,115200 and now I get console output. (Although not the complete log, it starts some time after boot)
The downstream kernel has some patches to simplify the UART naming. Please use ttyAMA0 or ttyS1. But it should provide the complete log.
The downstream kernel has some patches to simplify the UART naming. Please use ttyAMA0 or ttyS1. But it should provide the complete log.
its weird, it definitely misses off the start, when the terminal appears i’m getting junk being rx’d, i will try a different USB to serial just in case.
on an aside, ive just been looking around everywhere as ive been involved in another conversation about a usb library, it was suggested to use extcon to monitor gadget connection and disconnection events but the DWC2 driver has no references to the extcon API.
I’ve also ben studying the devicetree for the RPI device, and given that theres no schematics I’m assuming the phy is part of the DWC2 IP.
so is there any way of reading the VBUS state (level)? i eventually ended up reading a post in the kernel dev about the DWC2 driver by yourself!
It's possibly an issue with the baud rate.
Yes, i also assume the PHY is part of the DWC2 IP. Unfortunately we don't have a PHY driver yet, the devicetree fake this via non-op driver. Maybe this is the root cause of this issue. The BCM2835 datasheet provides a register description for MDIO access incl. VBUS interrupt.
I already started a template for a PHY driver: https://github.com/lategoodbye/rpi-zero/commits/bcm2835-dwc2-improvements but this actually needs more flesh
In case you need a datasheet for the DWC2 IP, try to google for Silicon Labs EZR32WG. It seems to use the same core.
It's possibly an issue with the baud rate.
Yes, i also assume the PHY is part of the DWC2 IP. Unfortunately we don't have a PHY driver yet, the devicetree fake this via non-op driver. Maybe this is the root cause of this issue. The BCM2835 provides a register description for MDIO access incl. VBUS interrupt.
I already started a template for a PHY driver: https://github.com/lategoodbye/rpi-zero/commits/bcm2835-dwc2-improvements but this actually needs more flesh
In case you need a datasheet for the DWC2 IP, try to google for Silicon Labs EZR32WG. It seems to use a the same core.
Awesome, I really appreciate your knowledge and help here. i’ll take a look at that document and have a read up on the VBUS interrupt to see if theres anything useful there, I’m trying to figure out a sane way of managing kernel rebuilds and transfer to the SD without network (in the case where i break the kernel) or taking apart my Pi Zero and removing the SD.
@lategoodbye i’ve had a quick skim of that datasheet and i concur its the same core. Having a read about as well, they (synopsis) also do phy ip, so we really have to guess at what the phy is.
The vbus sensing on that gecko part is done through a usb register, however, its located not in the same area of memory which means its not part of the core ip, its some other ip.
so without knowing how to read vbus, gadget mode is never going to work “as it should”. my hack is a fix for my particularl circumstance, but its not the correct fix. the correct fix requires the raspberry pi foundation to release some information on how to read the vbus information.
Is the MDIO register definition (p. 203 ff) in the BCM2835 datasheet not sufficient? What do you think, we need exactly? Sorry, i don't have in-depth USB knowledge.
Is the MDIO register definition (p. 203 ff) in the BCM2835 datasheet not sufficient? What do you think, we need exactly? Sorry, i don't have in-depth USB knowledge.
honestly, i didnt realise there was a datasheet for the SOC, I figured it was proprietary and that what had been released as source to the linux guys was all that was available from broadcom without NDA’s and such like.
I’ll go take a look.