rpi-4.4.y: Add USB gadget support to Pi Zero #1239

wants to merge 124 commits into


None yet
notro commented Dec 26, 2015

Build the dwc2 usb driver as a module to be able to use the USB On-the-Go functionality (supports both host and device/peripheral mode).
Add a specific Device Tree file for Pi Zero with dwc2 enabled in otg mode. Also fix the inverted ACT led.

This requires a change in the bootloader to give the Zero a dtb of it's own. Currently it uses this:

~$ sudo vcdbg log msg
001289.085: Loading 'bcm2708-rpi-b-plus.dtb' from SD card

This PR can be tested by setting the dtb manually in /boot/config.txt:



This will give a serial console on a connected computer:

sudo modprobe g_serial
sudo systemctl start getty@ttyGS0.service

On my Windows 8 computer, the USB serial port shows up as: ELMO GMAS (COM5)

Hotswapping to a keyboard works.

Controller fifo

The fifo assignment isn't optimal. This PR uses only 1024 of the 4080 available dynamic fifo entries:

[    8.779556] dwc2 20980000.usb: EPs: 8, dedicated fifos, 4080 entries in SPRAM

Currently assigned in Device Tree:

&usb {
    g-np-tx-fifo-size = <32>;
    g-rx-fifo-size = <256>;
    g-tx-fifo-size = <256 128 128 64 64 64 32>;

I don't know how to best distribute the remaining entries.


When I disconnect the keyboard, I get a warning:

[   93.639089] g_serial gadget: Gadget Serial v2.4
[   93.639120] g_serial gadget: g_serial ready
[   93.642186] dwc2 20980000.usb: bound driver g_serial
<connect computer>
[  101.877358] dwc2 20980000.usb: new device is high-speed
[  101.893135] dwc2 20980000.usb: new address 5
[  101.908600] g_serial gadget: high-speed config #2: CDC ACM config
<disconnect computer and connect keyboard>
[  158.138483] usb 1-1: new low-speed USB device number 2 using dwc2
[  158.355959] usb 1-1: New USB device found, idVendor=046d, idProduct=c313
[  158.356001] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[  158.356019] usb 1-1: Product: USB Multimedia Keyboard
[  158.356036] usb 1-1: Manufacturer: LITEON Technology
[  158.378340] input: LITEON Technology USB Multimedia Keyboard as /devices/platform/soc/20980000.usb/usb1/1-1/1-1:1.0/0003:046D:C313.0001/input/input0
[  158.441305] hid-generic 0003:046D:C313.0001: input,hidraw0: USB HID v1.10 Keyboard [LITEON Technology USB Multimedia Keyboard] on usb-20980000.usb-1/input0
[  158.454182] input: LITEON Technology USB Multimedia Keyboard as /devices/platform/soc/20980000.usb/usb1/1-1/1-1:1.1/0003:046D:C313.0002/input/input1
[  158.509498] hid-generic 0003:046D:C313.0002: input,hidraw1: USB HID v1.10 Device [LITEON Technology USB Multimedia Keyboard] on usb-20980000.usb-1/input1
<disconnect keyboard>
[  186.039220] usb 1-1: USB disconnect, device number 2
[  186.219073] ------------[ cut here ]------------
[  186.219241] WARNING: CPU: 0 PID: 67 at drivers/usb/dwc2/gadget.c:176 dwc2_hsotg_init_fifo+0x18c/0x1ac [dwc2]()
[  186.219258] Modules linked in: evdev usb_f_acm u_serial g_serial libcomposite cfg80211 rfkill dwc2 udc_core bcm2835_gpiomem bcm2835_wdt uio_pdrv_genirq uio i2c_dev snd_bcm2835 snd_pcm snd_timer snd fuse
[  186.219359] CPU: 0 PID: 67 Comm: kworker/u2:2 Not tainted 4.4.0-rc6+ #2
[  186.219371] Hardware name: BCM2708
[  186.219435] Workqueue: dwc2 dwc2_conn_id_status_change [dwc2]
[  186.219508] [<c0016ccc>] (unwind_backtrace) from [<c0013b3c>] (show_stack+0x20/0x24)
[  186.219548] [<c0013b3c>] (show_stack) from [<c032682c>] (dump_stack+0x20/0x28)
[  186.219584] [<c032682c>] (dump_stack) from [<c00222c8>] (warn_slowpath_common+0x8c/0xc4)
[  186.219613] [<c00222c8>] (warn_slowpath_common) from [<c00223bc>] (warn_slowpath_null+0x2c/0x34)
[  186.219697] [<c00223bc>] (warn_slowpath_null) from [<bf0ae6ac>] (dwc2_hsotg_init_fifo+0x18c/0x1ac [dwc2])
[  186.219831] [<bf0ae6ac>] (dwc2_hsotg_init_fifo [dwc2]) from [<bf0b04e4>] (dwc2_hsotg_core_init_disconnected+0x60/0x35c [dwc2])
[  186.219937] [<bf0b04e4>] (dwc2_hsotg_core_init_disconnected [dwc2]) from [<bf0a6824>] (dwc2_conn_id_status_change+0xe8/0x21c [dwc2])
[  186.220001] [<bf0a6824>] (dwc2_conn_id_status_change [dwc2]) from [<c0039c70>] (process_one_work+0x13c/0x490)
[  186.220029] [<c0039c70>] (process_one_work) from [<c003a13c>] (worker_thread+0x178/0x538)
[  186.220064] [<c003a13c>] (worker_thread) from [<c003fee8>] (kthread+0xe0/0xfc)
[  186.220097] [<c003fee8>] (kthread) from [<c000f788>] (ret_from_fork+0x14/0x2c)
[  186.220114] ---[ end trace 900ecfdec3d84a77 ]---
<connect computer>
[  189.792998] dwc2 20980000.usb: new device is high-speed
[  189.808759] dwc2 20980000.usb: new address 13
[  189.831151] g_serial gadget: high-speed config #2: CDC ACM config

Failed on 4.1

I also tried to do this PR on rpi-4.1.y but it didn't work and I didn't try to find out why. Got this error:

[    7.463269] dwc2 20980000.usb: no platform data or transceiver defined

This PR is based on work done by several people in #1212

@notro notro changed the title from Add USB gadget support to Pi Zero to rpi-4.4.y: Add USB gadget support to Pi Zero Dec 26, 2015
notro commented Dec 26, 2015

I only get the warning on keyboard disconnect if I have g_serial loaded.


Thanks. No objections. @pelwell ?

I did mention the inverted ACT led shortly before launch, but I think Mike liked the inverted, mostly on behaviour due to the absence of a PWR led, so it wasn't fixed then. I still think fixing it is probably the right thing to do, but we'll see what @pelwell and others think.


The ACT led is great how it works at the moment. Ideal as there is no PWR led. Please don't change it.

notro commented Dec 27, 2015

The ACT led is great how it works at the moment. Ideal as there is no PWR led.

In what way ideal, that it works excatly as a PWR led? The trigger can be changed to default-on to mimick a power led.

The problem with the inverted behaviour is that trigger none turns the led on and default-on turns it off. heartbeat gives a strange pulsing, and mmc0 is steady on apart from small "inverted blinks" when there's activity instead of a blink. cpu0 is also strange like mmc0.

There's a DT overlay parameter (act_led_activelow) to change the inversion, but I think that the expected behaviour should be the default.

Available triggers:

~$ cat /sys/class/leds/led0/trigger
none kbd-scrollock kbd-numlock kbd-capslock kbd-kanalock kbd-shiftlock kbd-altgrlock kbd-ctrllock kbd-altlock kbd-shiftllock kbd-shiftrlock kbd-ctrlllock kbd-ctrlrlock [mmc0] timer oneshot heartbeat backlight gpio cpu0 default-on input

Change trigger from boot with overlay parameter: act_led_trigger

Change trigger immediately:

~$ echo cpu0 | sudo tee /sys/class/leds/led0/trigger

I didn't say it works "exactly" as a PWR led, but it does provide a good indication the Pi zer0 is powered. It appears to default to ON indicating the Pi is powered and flashes off when the SD card is being accessed/written.

My Pi zero sat next to me is idle at the moment and the ACT led is lit showing it is being powered. If it is changed to behave like the earlier Pi's that have a seperate PWR led then it would be off and I would not know if it was powered on or not.

Or is this a case of it ain't broke fix it so it is, like the changes to the wifi made over the last few months.

clivem commented Dec 27, 2015

I think the original Zero LED behaviour should stay! I'm more interested in a visual indication that the Zero is being powered than its SDCARD is being accessed. What we have now is a good compromise, only having a single LED available to work with.

notro commented Dec 27, 2015

How about this change which will turn the led on, but keep the correct behaviour for the other triggers?

 &leds {
    act_led: act {
        label = "led0";
-       linux,default-trigger = "mmc0";
+       linux,default-trigger = "default-on";
        gpios = <&gpio 47 1>;

I agree that it's useful to have a power indicator.

Some thoughts on using the act led:
If I were to make a product based around a Zero, I would have turned the led on using dt-blob.bin, indicating that the board is powered (firmware loaded and sdcard funtional).
Then I would have used the trigger "timer" in the dtb causing the led to blink during kernel loading.
At last I would have turned the led steady on when the board was fully funtional.

This way I would now that if the led never lights, there's a power/sdcard problem. If it lights up I know the firmware is loaded.
If it doesn't start blinking, the kernel has failed to load.
If it stays blinking, I know that my app or whatever hasn't started succesfully.
There is also one more diagnostic step, and that would be to switch to the hearbeat pattern as the first thing when userspace loads. To show that systemd has started.
But I have never shipped a product so I don't know how useful it is to pinpoint these different stages for diagnostic purposes.

Steve Glendi... and others added some commits Feb 19, 2015
@popcornmix Steve Glendinning smsx95xx: fix crimes against truesize
smsc95xx is adjusting truesize when it shouldn't, and following a recent patch from Eric this is now triggering warnings.

This patch stops smsc95xx from changing truesize.

Signed-off-by: Steve Glendinning <steve.glendinning@smsc.com>
@popcornmix popcornmix vmstat: Workaround for issue where dirty page count goes negative 735dfca
@popcornmix popcornmix smsc95xx: Disable turbo mode by default 3c8125d
@notro @popcornmix notro irqchip: bcm2835: Add FIQ support
Add a duplicate irq range with an offset on the hwirq's so the
driver can detect that enable_fiq() is used.
Tested with downstream dwc_otg USB controller driver.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Acked-by: Stephen Warren <swarren@wwwdotorg.org>
@rtiemann @popcornmix rtiemann BCM2835_DT: Fix I2S register map 704ca73
@pelwell @popcornmix pelwell irq-bcm2836: Prevent spurious interrupts, and trap them early
The old arch-specific IRQ macros included a dsb to ensure the
write to clear the mailbox interrupt completed before returning
from the interrupt. The BCM2836 irqchip driver needs the same
precaution to avoid spurious interrupts.

Spurious interrupts are still possible for other reasons,
though, so trap them early.
@pelwell @popcornmix pelwell serial: 8250: Don't crash when nr_uarts is 0 8acf097
@notro @popcornmix notro irqchip: irq-bcm2835: Add 2836 FIQ support
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@notro @popcornmix notro pinctrl-bcm2835: Set base to 0 give expected gpio numbering
Signed-off-by: Noralf Tronnes <notro@tronnes.org>
@pelwell @popcornmix pelwell pinctrl-bcm2835: Fix interrupt handling for GPIOs 28-31 and 46-53
Contrary to the documentation, the BCM2835 GPIO controller actually has
four interrupt lines - one each for the three IRQ groups and one common. Rather
confusingly, the GPIO interrupt groups don't correspond directly with the GPIO
control banks. Instead, GPIOs 0-27 generate IRQ GPIO0, 28-45 GPIO1 and
46-53 GPIO2.

Awkwardly, the GPIOS for IRQ GPIO1 straddle two 32-entry GPIO banks, so it is
cleaner to split out a function to process the interrupts for a single GPIO

This bug has only just been observed because GPIOs above 27 can only be
accessed on an old Raspberry Pi with the optional P5 header fitted, where
the pins are often used for I2S instead.
@pelwell @popcornmix pelwell pinctrl-bcm2835: Only request the interrupts listed in the DTB
Although the GPIO controller can generate three interrupts (four counting
the common one), the device tree files currently only specify two. In the
absence of the third, simply don't register that interrupt (as opposed to
registering 0), which has the effect of making it impossible to generate
interrupts for GPIOs 46-53 which, since they share pins with the SD card
interface, is unlikely to be a problem.
@notro @popcornmix notro ARM: bcm2835: Set Serial number and Revision
The VideoCore bootloader passes in Serial number and
Revision number through Device Tree. Make these available to
userspace through /proc/cpuinfo.

Mainline status:

There is a commit in linux-next that standardize passing the serial
number through Device Tree (string: /serial-number):
ARM: 8355/1: arch: Show the serial number from devicetree in cpuinfo

There was an attempt to do the same with the revision number, but it
didn't get in:
[PATCH v2 1/2] arm: devtree: Set system_rev from DT revision

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@pelwell @popcornmix pelwell spi-bcm2835: Support pin groups other than 7-11
The spi-bcm2835 driver automatically uses GPIO chip-selects due to
some unreliability of the native ones. In doing so it chooses the
same pins as the native chip-selects would use, but the existing
code always uses pins 7 and 8, wherever the SPI function is mapped.

Search the pinctrl group assigned to the driver for pins that
correspond to native chip-selects, and use those for GPIO chip-

Signed-off-by: Phil Elwell <phil@raspberrypi.org>
@HiassofT @popcornmix HiassofT bcm2835-i2s: add 24bit support, update bclk_ratio to more correct values
Code ported from bcm2708-i2s driver in Raspberry Pi tree.

RPi commit 62c05a0 ("ASoC: BCM2708:
Add 24 bit support")

This adds 24 bit support to the I2S driver of the BCM2708.
Besides enabling the 24 bit flags, it includes two bug fixes:

MMAP is not supported. Claiming this leads to strange issues
when the format of driver and file do not match.

The datasheet states that the width extension bit should be set
for widths greater than 24, but greater or equal would be correct.
This follows from the definition of the width field.

Signed-off-by: Florian Meier <florian.meier@koalo.de>

RPi commit 3e8c672 ("bcm2708-i2s:
Update bclk_ratio to more correct values")

Discussion about blck_ratio affecting sound quality:

Signed-off-by: Matthias Reichl <hias@horus.com>
@HiassofT @popcornmix HiassofT bcm2835-i2s: get base address for DMA from devicetree
Code copied from spi-bcm2835. Get physical address from devicetree
instead of using hardcoded constant.

Signed-off-by: Matthias Reichl <hias@horus.com>
@HiassofT @popcornmix HiassofT bcm2835-i2s: setup clock only if CPU is clock master
Code ported from bcm2708-i2s driver in Raspberry Pi tree.

RPi commit c14827e ("bcm2708: Allow
option card devices to be configured via DT")

Original work by Zoltan Szenczi, committed to RPi tree by
Phil Elwell.

Signed-off-by: Matthias Reichl <hias@horus.com>
@HiassofT @popcornmix HiassofT bcm2835-i2s: Register PCM device
Code ported from bcm2708-i2s driver in Raspberry Pi tree.

RPi commit ba46b49 ("ASoC: Add
support for BCM2708")

This driver adds support for digital audio (I2S)
for the BCM2708 SoC that is used by the
Raspberry Pi. External audio codecs can be
connected to the Raspberry Pi via P5 header.

It relies on cyclic DMA engine support for BCM2708.

Signed-off-by: Florian Meier <florian.meier@koalo.de>

Signed-off-by: Matthias Reichl <hias@horus.com>
@HiassofT @popcornmix HiassofT bcm2835-i2s: Eliminate debugfs directory error
Code ported from bcm2708-i2s driver in Raspberry Pi tree.

RPi commit fd7d7a3 ("bcm2708:
Eliminate i2s debugfs directory error")

Qualify the two regmap ranges uses by bcm2708-i2s ('-i2s' and '-clk')
to avoid the name clash when registering debugfs entries.

Signed-off-by: Matthias Reichl <hias@horus.com>
@notro @popcornmix notro dmaengine: bcm2835: Add slave dma support
Add slave transfer capability to BCM2835 dmaengine driver.
This patch is pulled from the bcm2708-dmaengine driver in the
Raspberry Pi repo. The work was done by Gellert Weisz.

Tested using the bcm2835-mmc driver from the same repo.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@HiassofT @popcornmix HiassofT bcm2835-i2s: Enable MMAP support via a DT property
Code ported from bcm2708-i2s driver in Raspberry Pi tree.

RPi commit 7ee829f ("bcm2708-i2s:
Enable MMAP support via a DT property and overlay")

The i2s driver used to claim to support MMAP, but that feature was disabled
when some problems were found. Add the ability to enable this feature
through Device Tree, using the i2s-mmap overlay.

See: #1004

Signed-off-by: Matthias Reichl <hias@horus.com>
@notro @popcornmix notro dmaengine: bcm2835: set residue_granularity field
bcm2835-dma supports residue reporting at burst level but didn't report
this via the residue_granularity field.

Without this field set properly we get playback issues with I2S cards.

[by HiassofT, taken from bcm2708-dmaengine]
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@notro @popcornmix notro dmaengine: bcm2835: Load driver early and support legacy API
Load driver early since at least bcm2708_fb doesn't support deferred
probing and even if it did, we don't want the video driver deferred.
Support the legacy DMA API which is needed by bcm2708_fb.
Don't mask out channel 2.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@notro @popcornmix notro bcm2835: Add support for uart1
This is a hack until a proper solution is agreed upon.
Martin Sperl is doing some work in this area.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@HiassofT @popcornmix HiassofT bcm2835-dma: Fix dreq not set for slave transfers
Set dreq to slave_id if it is not set like in bcm2708-dmaengine.
@HiassofT @popcornmix HiassofT bcm2835-dma: Limit cyclic transfers on lite channels to 32k
Transfers larger than 32k cause repeated clicking with I2S soundcards.
The exact reason is yet unknown, so limit to 32k as bcm2708-dmaengine
did as an intermediate fix.
@popcornmix popcornmix Main bcm2708/bcm2709 linux port
Signed-off-by: popcornmix <popcornmix@gmail.com>
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@notro @popcornmix notro firmware: bcm2835: Add missing property tags
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@popcornmix popcornmix Add dwc_otg driver
Signed-off-by: popcornmix <popcornmix@gmail.com>

usb: dwc: fix lockdep false positive

Signed-off-by: Kari Suvanto <karis79@gmail.com>

usb: dwc: fix inconsistent lock state

Signed-off-by: Kari Suvanto <karis79@gmail.com>

Add FIQ patch to dwc_otg driver. Enable with dwc_otg.fiq_fix_enable=1. Should give about 10% more ARM performance.
Thanks to Gordon and Costas

Avoid dynamic memory allocation for channel lock in USB driver. Thanks ddv2005.

Add NAK holdoff scheme. Enabled by default, disable with dwc_otg.nak_holdoff_enable=0. Thanks gsh

Make sure we wait for the reset to finish

dwc_otg: fix bug in dwc_otg_hcd.c resulting in silent kernel
	 memory corruption, escalating to OOPS under high USB load.

dwc_otg: Fix unsafe access of QTD during URB enqueue

In dwc_otg_hcd_urb_enqueue during qtd creation, it was possible that the
transaction could complete almost immediately after the qtd was assigned
to a host channel during URB enqueue, which meant the qtd pointer was no
longer valid having been completed and removed. Usually, this resulted in
an OOPS during URB submission. By predetermining whether transactions
need to be queued or not, this unsafe pointer access is avoided.

This bug was only evident on the Pi model A where a device was attached
that had no periodic endpoints (e.g. USB pendrive or some wlan devices).

dwc_otg: Fix incorrect URB allocation error handling

If the memory allocation for a dwc_otg_urb failed, the kernel would OOPS
because for some reason a member of the *unallocated* struct was set to
zero. Error handling changed to fail correctly.

dwc_otg: fix potential use-after-free case in interrupt handler

If a transaction had previously aborted, certain interrupts are
enabled to track error counts and reset where necessary. On IN
endpoints the host generates an ACK interrupt near-simultaneously
with completion of transfer. In the case where this transfer had
previously had an error, this results in a use-after-free on
the QTD memory space with a 1-byte length being overwritten to

dwc_otg: add handling of SPLIT transaction data toggle errors

Previously a data toggle error on packets from a USB1.1 device behind
a TT would result in the Pi locking up as the driver never handled
the associated interrupt. Patch adds basic retry mechanism and
interrupt acknowledgement to cater for either a chance toggle error or
for devices that have a broken initial toggle state (FT8U232/FT232BM).

dwc_otg: implement tasklet for returning URBs to usbcore hcd layer

The dwc_otg driver interrupt handler for transfer completion will spend
a very long time with interrupts disabled when a URB is completed -
this is because usb_hcd_giveback_urb is called from within the handler
which for a USB device driver with complicated processing (e.g. webcam)
will take an exorbitant amount of time to complete. This results in
missed completion interrupts for other USB packets which lead to them
being dropped due to microframe overruns.

This patch splits returning the URB to the usb hcd layer into a
high-priority tasklet. This will have most benefit for isochronous IN
transfers but will also have incidental benefit where multiple periodic
devices are active at once.

dwc_otg: fix NAK holdoff and allow on split transactions only

This corrects a bug where if a single active non-periodic endpoint
had at least one transaction in its qh, on frnum == MAX_FRNUM the qh
would get skipped and never get queued again. This would result in
a silent device until error detection (automatic or otherwise) would
either reset the device or flush and requeue the URBs.

Additionally the NAK holdoff was enabled for all transactions - this
would potentially stall a HS endpoint for 1ms if a previous error state
enabled this interrupt and the next response was a NAK. Fix so that
only split transactions get held off.

dwc_otg: Call usb_hcd_unlink_urb_from_ep with lock held in completion handler

usb_hcd_unlink_urb_from_ep must be called with the HCD lock held.  Calling it
asynchronously in the tasklet was not safe (regression in

This change unlinks it from the endpoint prior to queueing it for handling in
the tasklet, and also adds a check to ensure the urb is OK to be unlinked
before doing so.

NULL pointer dereference kernel oopses had been observed in usb_hcd_giveback_urb
when a USB device was unplugged/replugged during data transfer.  This effect
was reproduced using automated USB port power control, hundreds of replug
events were performed during active transfers to confirm that the problem was

USB fix using a FIQ to implement split transactions

This commit adds a FIQ implementaion that schedules
the split transactions using a FIQ so we don't get
held off by the interrupt latency of Linux

dwc_otg: fix device attributes and avoid kernel warnings on boot

dcw_otg: avoid logging function that can cause panics

See: raspberrypi/firmware#21
Thanks to cleverca22 for fix

dwc_otg: mask correct interrupts after transaction error recovery

The dwc_otg driver will unmask certain interrupts on a transaction
that previously halted in the error state in order to reset the
QTD error count. The various fine-grained interrupt handlers do not
consider that other interrupts besides themselves were unmasked.

By disabling the two other interrupts only ever enabled in DMA mode
for this purpose, we can avoid unnecessary function calls in the
IRQ handler. This will also prevent an unneccesary FIQ interrupt
from being generated if the FIQ is enabled.

dwc_otg: fiq: prevent FIQ thrash and incorrect state passing to IRQ

In the case of a transaction to a device that had previously aborted
due to an error, several interrupts are enabled to reset the error
count when a device responds. This has the side-effect of making the
FIQ thrash because the hardware will generate multiple instances of
a NAK on an IN bulk/interrupt endpoint and multiple instances of ACK
on an OUT bulk/interrupt endpoint. Make the FIQ mask and clear the
associated interrupts.

Additionally, on non-split transactions make sure that only unmasked
interrupts are cleared. This caused a hard-to-trigger but serious
race condition when you had the combination of an endpoint awaiting
error recovery and a transaction completed on an endpoint - due to
the sequencing and timing of interrupts generated by the dwc_otg core,
it was possible to confuse the IRQ handler.

Fix function tracing

dwc_otg: whitespace cleanup in dwc_otg_urb_enqueue

dwc_otg: prevent OOPSes during device disconnects

The dwc_otg_urb_enqueue function is thread-unsafe. In particular the
access of urb->hcpriv, usb_hcd_link_urb_to_ep, dwc_otg_urb->qtd and
friends does not occur within a critical section and so if a device
was unplugged during activity there was a high chance that the
usbcore hub_thread would try to disable the endpoint with partially-
formed entries in the URB queue. This would result in BUG() or null
pointer dereferences.

Fix so that access of urb->hcpriv, enqueuing to the hardware and
adding to usbcore endpoint URB lists is contained within a single
critical section.

dwc_otg: prevent BUG() in TT allocation if hub address is > 16

A fixed-size array is used to track TT allocation. This was
previously set to 16 which caused a crash because
dwc_otg_hcd_allocate_port would read past the end of the array.

This was hit if a hub was plugged in which enumerated as addr > 16,
due to previous device resets or unplugs.

Also add #ifdef FIQ_DEBUG around hcd->hub_port_alloc[], which grows
to a large size if 128 hub addresses are supported. This field is
for debug only for tracking which frame an allocate happened in.

dwc_otg: make channel halts with unknown state less damaging

If the IRQ received a channel halt interrupt through the FIQ
with no other bits set, the IRQ would not release the host
channel and never complete the URB.

Add catchall handling to treat as a transaction error and retry.

dwc_otg: fiq_split: use TTs with more granularity

This fixes certain issues with split transaction scheduling.

- Isochronous multi-packet OUT transactions now hog the TT until
  they are completed - this prevents hubs aborting transactions
  if they get a periodic start-split out-of-order
- Don't perform TT allocation on non-periodic endpoints - this
  allows simultaneous use of the TT's bulk/control and periodic
  transaction buffers

This commit will mainly affect USB audio playback.

dwc_otg: fix potential sleep while atomic during urb enqueue

Fixes a regression introduced with eb1b482. Kmalloc called from
dwc_otg_hcd_qtd_add / dwc_otg_hcd_qtd_create did not always have
the GPF_ATOMIC flag set. Force this flag when inside the larger
critical section.

dwc_otg: make fiq_split_enable imply fiq_fix_enable

Failing to set up the FIQ correctly would result in
"IRQ 32: nobody cared" errors in dmesg.

dwc_otg: prevent crashes on host port disconnects

Fix several issues resulting in crashes or inconsistent state
if a Model A root port was disconnected.

- Clean up queue heads properly in kill_urbs_in_qh_list by
  removing the empty QHs from the schedule lists
- Set the halt status properly to prevent IRQ handlers from
  using freed memory
- Add fiq_split related cleanup for saved registers
- Make microframe scheduling reclaim host channels if
  active during a disconnect
- Abort URBs with -ESHUTDOWN status response, informing
  device drivers so they respond in a more correct fashion
  and don't try to resubmit URBs
- Prevent IRQ handlers from attempting to handle channel
  interrupts if the associated URB was dequeued (and the
  driver state was cleared)

dwc_otg: prevent leaking URBs during enqueue

A dwc_otg_urb would get leaked if the HCD enqueue function
failed for any reason. Free the URB at the appropriate points.

dwc_otg: Enable NAK holdoff for control split transactions

Certain low-speed devices take a very long time to complete a
data or status stage of a control transaction, producing NAK
responses until they complete internal processing - the USB2.0
spec limit is up to 500mS. This causes the same type of interrupt
storm as seen with USB-serial dongles prior to c8edb23.

In certain circumstances, usually while booting, this interrupt
storm could cause SD card timeouts.

dwc_otg: Fix for occasional lockup on boot when doing a USB reset

dwc_otg: Don't issue traffic to LS devices in FS mode

Issuing low-speed packets when the root port is in full-speed mode
causes the root port to stop responding. Explicitly fail when
enqueuing URBs to a LS endpoint on a FS bus.

Fix ARM architecture issue with local_irq_restore()

If local_fiq_enable() is called before a local_irq_restore(flags) where
the flags variable has the F bit set, the FIQ will be erroneously disabled.

Fixup arch_local_irq_restore to avoid trampling the F bit in CPSR.

Also fix some of the hacks previously implemented for previous dwc_otg

dwc_otg: fiq_fsm: Base commit for driver rewrite

This commit removes the previous FIQ fixes entirely and adds fiq_fsm.

This rewrite features much more complete support for split transactions
and takes into account several OTG hardware bugs. High-speed
isochronous transactions are also capable of being performed by fiq_fsm.

All driver options have been removed and replaced with:
  - dwc_otg.fiq_enable (bool)
  - dwc_otg.fiq_fsm_enable (bool)
  - dwc_otg.fiq_fsm_mask (bitmask)
  - dwc_otg.nak_holdoff (unsigned int)

Defaults are specified such that fiq_fsm behaves similarly to the
previously implemented FIQ fixes.

fiq_fsm: Push error recovery into the FIQ when fiq_fsm is used

If the transfer associated with a QTD failed due to a bus error, the HCD
would retry the transfer up to 3 times (implementing the USB2.0
three-strikes retry in software).

Due to the masking mechanism used by fiq_fsm, it is only possible to pass
a single interrupt through to the HCD per-transfer.

In this instance host channels would fall off the radar because the error
reset would function, but the subsequent channel halt would be lost.

Push the error count reset into the FIQ handler.

fiq_fsm: Implement timeout mechanism

For full-speed endpoints with a large packet size, interrupt latency
runs the risk of the FIQ starting a transaction too late in a full-speed
frame. If the device is still transmitting data when EOF2 for the
downstream frame occurs, the hub will disable the port. This change is
not reflected in the hub status endpoint and the device becomes

Prevent high-bandwidth transactions from being started too late in a
frame. The mechanism is not guaranteed: a combination of bit stuffing
and hub latency may still result in a device overrunning.

fiq_fsm: fix bounce buffer utilisation for Isochronous OUT

Multi-packet isochronous OUT transactions were subject to a few bounday
bugs. Fix them.

Audio playback is now much more robust: however, an issue stands with
devices that have adaptive sinks - ALSA plays samples too fast.

dwc_otg: Return full-speed frame numbers in HS mode

The frame counter increments on every *microframe* in high-speed mode.
Most device drivers expect this number to be in full-speed frames - this
caused considerable confusion to e.g. snd_usb_audio which uses the
frame counter to estimate the number of samples played.

fiq_fsm: save PID on completion of interrupt OUT transfers

Also add edge case handling for interrupt transports.

Note that for periodic split IN, data toggles are unimplemented in the
OTG host hardware - it unconditionally accepts any PID.

fiq_fsm: add missing case for fiq_fsm_tt_in_use()

Certain combinations of bitrate and endpoint activity could
result in a periodic transaction erroneously getting started
while the previous Isochronous OUT was still active.

fiq_fsm: clear hcintmsk for aborted transactions

Prevents the FIQ from erroneously handling interrupts
on a timed out channel.

fiq_fsm: enable by default

fiq_fsm: fix dequeues for non-periodic split transactions

If a dequeue happened between the SSPLIT and CSPLIT phases of the
transaction, the HCD would never receive an interrupt.

fiq_fsm: Disable by default

fiq_fsm: Handle HC babble errors

The HCTSIZ transfer size field raises a babble interrupt if
the counter wraps. Handle the resulting interrupt in this case.

dwc_otg: fix interrupt registration for fiq_enable=0

Additionally make the module parameter conditional for wherever
hcd->fiq_state is touched.

fiq_fsm: Enable by default

dwc_otg: Fix various issues with root port and transaction errors

Process the host port interrupts correctly (and don't trample them).
Root port hotplug now functional again.

Fix a few thinkos with the transaction error passthrough for fiq_fsm.

fiq_fsm: Implement hack for Split Interrupt transactions

Hubs aren't too picky about which endpoint we send Control type split
transactions to. By treating Interrupt transfers as Control, it is
possible to use the non-periodic queue in the OTG core as well as the
non-periodic FIFOs in the hub itself. This massively reduces the
microframe exclusivity/contention that periodic split transactions
otherwise have to enforce.

It goes without saying that this is a fairly egregious USB specification
violation, but it works.

Original idea by Hans Petter Selasky @ FreeBSD.org.

dwc_otg: FIQ support on SMP. Set up FIQ stack and handler on Core 0 only.

dwc_otg: introduce fiq_fsm_spin(un|)lock()

SMP safety for the FIQ relies on register read-modify write cycles being
completed in the correct order. Several places in the DWC code modify
registers also touched by the FIQ. Protect these by a bare-bones lock

This also makes it possible to run the FIQ and IRQ handlers on different

fiq_fsm: fix build on bcm2708 and bcm2709 platforms

dwc_otg: put some barriers back where they should be for UP

bcm2709/dwc_otg: Setup FIQ on core 1 if >1 core active

dwc_otg: fixup read-modify-write in critical paths

Be more careful about read-modify-write on registers that the FIQ
also touches.

Guard fiq_fsm_spin_lock with fiq_enable check

fiq_fsm: Falling out of the state machine isn't fatal

This edge case can be hit if the port is disabled while the FIQ is
in the middle of a transaction. Make the effects less severe.

Also get rid of the useless return value.

squash: dwc_otg: Allow to build without SMP

usb: core: make overcurrent messages more prominent

Hub overcurrent messages are more serious than "debug". Increase loglevel.

usb: dwc_otg: Don't use dma_to_virt()

Commit 6ce0d20 changes dma_to_virt() which breaks this driver.
Open code the old dma_to_virt() implementation to work around this.

Limit the use of __bus_to_virt() to cases where transfer_buffer_length
is set and transfer_buffer is not set. This is done to increase the
chance that this driver will also work on ARCH_BCM2835.

transfer_buffer should not be NULL if the length is set, but the
comment in the code indicates that there are situations where this
might happen. drivers/usb/isp1760/isp1760-hcd.c also has a similar
comment pointing to a possible: 'usb storage / SCSI bug'.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

dwc_otg: Fix crash when fiq_enable=0

dwc_otg: fiq_fsm: Make high-speed isochronous strided transfers work properly

Certain low-bandwidth high-speed USB devices (specialist audio devices,
compressed-frame webcams) have packet intervals > 1 microframe.

Stride these transfers in the FIQ by using the start-of-frame interrupt
to restart the channel at the right time.

dwc_otg: Force host mode to fix incorrect compute module boards

dwc_otg: Add ARCH_BCM2835 support

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

dwc_otg: Simplify FIQ irq number code

Dropping ATAGS means we can simplify the FIQ irq number code.
Also add error checking on the returned irq number.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

dwc_otg: Remove duplicate gadget probe/unregister function
@popcornmix popcornmix squash: include ARCH_BCM2708 / ARCH_BCM2709 f2d3c1b
@koalo @popcornmix koalo dmaengine: Add support for BCM2708
Add support for DMA controller of BCM2708 as used in the Raspberry Pi.
Currently it only supports cyclic DMA.

Signed-off-by: Florian Meier <florian.meier@koalo.de>

dmaengine: expand functionality by supporting scatter/gather transfers sdhci-bcm2708 and dma.c: fix for LITE channels

DMA: fix cyclic LITE length overflow bug

dmaengine: bcm2708: Remove chancnt affectations

Mirror bcm2835-dma.c commit 9eba553:
chancnt is already filled by dma_async_device_register, which uses the channel
list to know how much channels there is.

Since it's already filled, we can safely remove it from the drivers' probe

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

dmaengine: bcm2708: overwrite dreq only if it is not set

dreq is set when the DMA channel is fetched from Device Tree.
slave_id is set using dmaengine_slave_config().
Only overwrite dreq with slave_id if it is not set.

dreq/slave_id in the cyclic DMA case is not touched, because I don't
have hardware to test with.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

dmaengine: bcm2708: do device registration in the board file

Don't register the device in the driver. Do it in the board file.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

dmaengine: bcm2708: don't restrict DT support to ARCH_BCM2835

Both ARCH_BCM2835 and ARCH_BCM270x are built with OF now.
Add Device Tree support to the non ARCH_BCM2835 case.
Use the same driver name regardless of architecture.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

BCM270x_DT: add bcm2835-dma entry

Add Device Tree entry for bcm2835-dma.
The entry doesn't contain any resources since they are handled
by the arch/arm/mach-bcm270x/dma.c driver.
In non-DT mode, don't add the device in the board file.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

bcm2708-dmaengine: Add debug options

BCM270x: Add memory and irq resources to dmaengine device and DT

Prepare for merging of the legacy DMA API arch driver dma.c
with bcm2708-dmaengine by adding memory and irq resources both
to platform file device and Device Tree node.
Don't use BCM_DMAMAN_DRIVER_NAME so we don't have to include mach/dma.h

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

dmaengine: bcm2708: Merge with arch dma.c driver and disable dma.c

Merge the legacy DMA API driver with bcm2708-dmaengine.
This is done so we can use bcm2708_fb on ARCH_BCM2835 (mailbox
driver is also needed).

Changes to the dma.c code:
- Use BIT() macro.
- Cutdown some comments to one line.
- Add mutex to vc_dmaman and use this, since the dev lock is locked
  during probing of the engine part.
- Add global g_dmaman variable since drvdata is used by the engine part.
- Restructure for readability:
- Restructure bcm_dma_chan_alloc() to simplify error handling.
- Use device irq resources instead of hardcoded bcm_dma_irqs table.
- Remove dev_dmaman_register() and code it directly.
- Remove dev_dmaman_deregister() and code it directly.
- Simplify bcm_dmaman_probe() using devm_* functions.
- Get dmachans from DT if available.
- Keep 'dma.dmachans' module argument name for backwards compatibility.

Make it available on ARCH_BCM2835 as well.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

dmaengine: bcm2708: set residue_granularity field

bcm2708-dmaengine supports residue reporting at burst level
but didn't report this via the residue_granularity field.

Without this field set properly we get playback issues with I2S cards.

dmaengine: bcm2708-dmaengine: Fix memory leak when stopping a running transfer

bcm2708-dmaengine: Use more DMA channels (but not 12)

1) Only the bcm2708_fb drivers uses the legacy DMA API, and
it requires a BULK-capable channel, so all other types
(FAST, NORMAL and LITE) can be made available to the regular

2) DMA channels 11-14 share an interrupt. The driver can't
handle this, so don't use channels 12-14 (12 was used, probably
because it appears to have an interrupt, but in reality that
interrupt is for activity on ANY channel). This may explain
a lockup encountered when running out of DMA channels.

The combined effect of this patch is to leave 7 DMA channels
available + channel 0 for bcm2708_fb via the legacy API.

See: raspberrypi#1110

dmaengine: bcm2708: Make legacy API available for bcm2835-dma

bcm2708_fb uses the legacy DMA API, so in order to start using
bcm2835-dma, bcm2835-dma has to support the legacy API. Make this
possible by exporting bcm_dmaman_probe() and bcm_dmaman_remove().

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

dmaengine: bcm2708: Change DT compatible string

Both bcm2835-dma and bcm2708-dmaengine have the same compatible string.
So change compatible to "brcm,bcm2708-dma".

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

dmaengine: bcm2708: Remove driver but keep legacy API

Dropping non-DT support means we don't need this driver,
but we still need the legacy DMA API.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@popcornmix popcornmix bcm2708 framebuffer driver
Signed-off-by: popcornmix <popcornmix@gmail.com>

bcm2708_fb : Implement blanking support using the mailbox property interface

bcm2708_fb: Add pan and vsync controls

bcm2708_fb: DMA acceleration for fb_copyarea

Based on http://www.raspberrypi.org/phpBB3/viewtopic.php?p=62425#p62425
Also used Simon's dmaer_master module as a reference for tweaking DMA
settings for better performance.

For now busylooping only. IRQ support might be added later.
With non-overclocked Raspberry Pi, the performance is ~360 MB/s
for simple copy or ~260 MB/s for two-pass copy (used when dragging
windows to the right).

In the case of using DMA channel 0, the performance improves
to ~440 MB/s.

For comparison, VFP optimized CPU copy can only do ~114 MB/s in
the same conditions (hindered by reading uncached source buffer).

Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>

bcm2708_fb: report number of dma copies

Add a counter (exported via debugfs) reporting the
number of dma copies that the framebuffer driver
has done, in order to help evaluate different
optimization strategies.

Signed-off-by: Luke Diamand <luked@broadcom.com>

bcm2708_fb: use IRQ for DMA copies

The copyarea ioctl() uses DMA to speed things along. This
was busy-waiting for completion. This change supports using
an interrupt instead for larger transfers. For small
transfers, busy-waiting is still likely to be faster.

Signed-off-by: Luke Diamand <luke@diamand.org>

bcm2708: Make ioctl logging quieter

video: fbdev: bcm2708_fb: Don't panic on error

No need to panic the kernel if the video driver fails.
Just print a message and return an error.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

fbdev: bcm2708_fb: Add ARCH_BCM2835 support

Add Device Tree support.
Pass the device to dma_alloc_coherent() in order to get the
correct bus address on ARCH_BCM2835.
Use the new DMA legacy API header file.
Including <mach/platform.h> is not necessary.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

BCM270x_DT: Add bcm2708-fb device

Add bcm2708-fb to Device Tree and don't add the
platform device when booting in DT mode.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@weiszg @popcornmix weiszg MMC: added alternative MMC driver
mmc: Disable CMD23 transfers on all cards

Pending wire-level investigation of these types of transfers
and associated errors on bcm2835-mmc, disable for now. Fallback of
CMD18/CMD25 transfers will be used automatically by the MMC layer.

Reported/Tested-by: Gellert Weisz <gellert@raspberrypi.org>

mmc: bcm2835-mmc: enable DT support for all architectures

Both ARCH_BCM2835 and ARCH_BCM270x are built with OF now.
Enable Device Tree support for all architectures.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

mmc: bcm2835-mmc: fix probe error handling

Probe error handling is broken in several places.
Simplify error handling by using device managed functions.
Replace pr_{err,info} with dev_{err,info}.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

bcm2835-mmc: Add locks when accessing sdhost registers

bcm2835-mmc: Add range of debug options for slowing things down

bcm2835-mmc: Add option to disable some delays

bcm2835-mmc: Add option to disable MMC_QUIRK_BLK_NO_CMD23

bcm2835-mmc: Default to disabling MMC_QUIRK_BLK_NO_CMD23

bcm2835-mmc: Adding overclocking option

Allow a different clock speed to be substitued for a requested 50MHz.
This option is exposed using the "overclock_50" DT parameter.
Note that the mmc interface is restricted to EVEN integer divisions of
250MHz, and the highest sensible option is 63 (250/4 = 62.5), the
next being 125 (250/2) which is much too high.

Use at your own risk.

bcm2835-mmc: Round up the overclock, so 62 works for 62.5Mhz

Also only warn once for each overclock setting.

mmc: bcm2835-mmc: Make available on ARCH_BCM2835

Make the bcm2835-mmc driver available for use on ARCH_BCM2835.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

BCM270x_DT: add bcm2835-mmc entry

Add Device Tree entry for bcm2835-mmc.
In non-DT mode, don't add the device in the board file.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

bcm2835-mmc: Don't overwrite MMC capabilities from DT
@pelwell @popcornmix pelwell Add blk_pos parameter to mmc multi_io_quirk callback 09bf5ed
@popcornmix popcornmix cma: Add vc_cma driver to enable use of CMA
Signed-off-by: popcornmix <popcornmix@gmail.com>

vc_cma: Make the vc_cma area the default contiguous DMA area

vc_cma: Provide empty functions when module is not built

Providing empty functions saves the users from guarding the
function call with an #if clause.
Move __init markings from prototypes to functions.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@pelwell @popcornmix pelwell Adding bcm2835-sdhost driver, and an overlay to enable it
BCM2835 has two SD card interfaces. This driver uses the other one.

bcm2835-sdhost: Error handling fix, and code clarification

bcm2835-sdhost: Adding overclocking option

Allow a different clock speed to be substitued for a requested 50MHz.
This option is exposed using the "overclock_50" DT parameter.
Note that the sdhost interface is restricted to integer divisions of
core_freq, and the highest sensible option for a core_freq of 250MHz
is 84 (250/3 = 83.3MHz), the next being 125 (250/2) which is much too

Use at your own risk.

bcm2835-sdhost: Round up the overclock, so 62 works for 62.5Mhz

Also only warn once for each overclock setting.

bcm2835-sdhost: Improve error handling and recovery

1) Expose the hw_reset method to the MMC framework, removing many
   internal calls by the driver.

2) Reduce overclock setting on error.

3) Increase timeout to cope with high capacity cards.

4) Add properties and parameters to control pio_limit and debug.

5) Reduce messages at probe time.

bcm2835-sdhost: Further improve overclock back-off

bcm2835-sdhost: Clear HBLC for PIO mode

Also update pio_limit default in overlay README.

bcm2835-sdhost: Add the ERASE capability

See: raspberrypi#1076

bcm2835-sdhost: Ignore CRC7 for MMC CMD1

It seems that the sdhost interface returns CRC7 errors for CMD1,
which is the MMC-specific SEND_OP_COND. Returning these errors to
the MMC layer causes a downward spiral, but ignoring them seems
to be harmless.

bcm2835-mmc/sdhost: Remove ARCH_BCM2835 differences

The bcm2835-mmc driver (and -sdhost driver that copied from it)
contains code to handle SDIO interrupts in a threaded interrupt
handler rather than waking the MMC framework thread. The change
follows a patch from Russell King that adds the facility as the
preferred way of working.

However, the new code path is only present in ARCH_BCM2835
builds, which I have taken to be a way of testing the waters
rather than making the change across the board; I can't see
any technical reason why it wouldn't be enabled for MACH_BCM270X
builds. So this patch standardises on the ARCH_BCM2835 code,
removing the old code paths.

bcm2835-sdhost: Don't log timeout errors unless debug=1

The MMC card-discovery process generates timeouts. This is
expected behaviour, so reporting it to the user serves no purpose.
Suppress the reporting of timeout errors unless the debug flag
is on.
@popcornmix popcornmix bcm2708: alsa sound driver
Signed-off-by: popcornmix <popcornmix@gmail.com>

alsa: add mmap support and some cleanups to bcm2835 ALSA driver

snd-bcm2835: Add support for spdif/hdmi passthrough

This adds a dedicated subdevice which can be used for passthrough of non-audio
formats (ie encoded a52) through the hdmi audio link. In addition to this
driver extension an appropriate card config is required to make alsa-lib
support the AES parameters for this device.

snd-bcm2708: Add mutex, improve logging

Fix for ALSA driver crash

Avoids an issue when closing and opening vchiq where a message can arrive before service handle has been written

alsa: reduce severity of expected warning message

snd-bcm2708: Fix dmesg spam for non-error case

alsa: Ensure mutexes are released through error paths

alsa: Make interrupted close paths quieter

BCM270x: Add onboard sound device to Device Tree

Add Device Tree support to alsa driver.
Add device to Device Tree.
Don't add platform devices when booting in DT mode.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@popcornmix popcornmix bcm2708 vchiq driver
Signed-off-by: popcornmix <popcornmix@gmail.com>

vchiq: create_pagelist copes with vmalloc memory

Signed-off-by: Daniel Stone <daniels@collabora.com>

vchiq: fix the shim message release

Signed-off-by: Daniel Stone <daniels@collabora.com>

vchiq: export additional symbols

Signed-off-by: Daniel Stone <daniels@collabora.com>

VCHIQ: Make service closure fully synchronous (drv)

This is one half of a two-part patch, the other half of which is to
the vchiq_lib user library. With these patches, calls to
vchiq_close_service and vchiq_remove_service won't return until any
associated callbacks have been delivered to the callback thread.

VCHIQ: Add per-service tracing

The new service option VCHIQ_SERVICE_OPTION_TRACE is a boolean that
toggles tracing for the specified service.

This commit also introduces vchi_service_set_option and the associated

vchiq: Make the synchronous-CLOSE logic more tolerant

vchiq: Move logging control into debugfs

vchiq: Take care of a corner case tickled by VCSM

Closing a connection that isn't fully open requires care, since one
side does not know the other side's port number. Code was present to
handle the case where a CLOSE is sent immediately after an OPEN, i.e.
before the OPENACK has been received, but this was incorrectly being
used when an OPEN from a client using port 0 was rejected.

(In the observed failure, the host was attempting to use the VCSM
service, which isn't present in the 'cutdown' firmware. The failure
was intermittent because sometimes the keepalive service would
grab port 0.)

This case can be distinguished because the client's remoteport will
still be VCHIQ_PORT_FREE, and the srvstate will be OPENING. Either
condition is sufficient to differentiate it from the special case
described above.

vchiq: Avoid high load when blocked and unkillable

vchiq: Include SIGSTOP and SIGCONT in list of signals not-masked by vchiq to allow gdb to work

vchiq_arm: Complete support for SYNCHRONOUS mode

vchiq: Remove inline from suspend/resume

vchiq: Allocation does not need to be atomic

vchiq: Fix wrong condition check

The log level is checked from within the log call. Remove the check in the call.

Signed-off-by: Pranith Kumar <bobby.prani@gmail.com>

BCM270x: Add vchiq device to platform file and Device Tree

Prepare to turn the vchiq module into a driver.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

bcm2708: vchiq: Add Device Tree support

Turn vchiq into a driver and stop hardcoding resources.
Use devm_* functions in probe path to simplify cleanup.
A global variable is used to hold the register address. This is done
to keep this patch as small as possible.
Also make available on ARCH_BCM2835.
Based on work by Lubomir Rintel.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

vchiq: Change logging level for inbound data

vchiq_arm: Two cacheing fixes

1) Make fragment size vary with cache line size
Without this patch, non-cache-line-aligned transfers may corrupt
(or be corrupted by) adjacent data structures.

Both ARM and VC need to be updated to enable this feature. This is
ensured by having the loader apply a new DT parameter -
cache-line-size. The existence of this parameter guarantees that the
kernel is capable, and the parameter will only be modified from the
safe default if the loader is capable.

2) Flush/invalidate vmalloc'd memory, and invalidate after reads

vchiq: fix NULL pointer dereference when closing driver

The following code run as root will cause a null pointer dereference oops:

        int fd = open("/dev/vc-cma", O_RDONLY);
        if (fd < 0)
                err(1, "open failed");

[ 1704.877721] Unable to handle kernel NULL pointer dereference at virtual address 00000000
[ 1704.877725] pgd = b899c000
[ 1704.877736] [00000000] *pgd=37fab831, *pte=00000000, *ppte=00000000
[ 1704.877748] Internal error: Oops: 817 [#1] PREEMPT SMP ARM
[ 1704.877765] Modules linked in: evdev i2c_bcm2708 uio_pdrv_genirq uio
[ 1704.877774] CPU: 2 PID: 3656 Comm: stress-ng-fstat Not tainted 3.19.1-12-generic-bcm2709 #12-Ubuntu
[ 1704.877777] Hardware name: BCM2709
[ 1704.877783] task: b8ab9b00 ti: b7e68000 task.ti: b7e68000
[ 1704.877798] PC is at __down_interruptible+0x50/0xec
[ 1704.877806] LR is at down_interruptible+0x5c/0x68
[ 1704.877813] pc : [<80630ee8>]    lr : [<800704b0>]    psr: 60080093
sp : b7e69e50  ip : b7e69e88  fp : b7e69e84
[ 1704.877817] r10: b88123c8  r9 : 00000010  r8 : 00000001
[ 1704.877822] r7 : b8ab9b00  r6 : 7fffffff  r5 : 80a1cc34  r4 : 80a1cc34
[ 1704.877826] r3 : b7e69e50  r2 : 00000000  r1 : 00000000  r0 : 80a1cc34
[ 1704.877833] Flags: nZCv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  Segment user
[ 1704.877838] Control: 10c5387d  Table: 3899c06a  DAC: 00000015
[ 1704.877843] Process do-oops (pid: 3656, stack limit = 0xb7e68238)
[ 1704.877848] Stack: (0xb7e69e50 to 0xb7e6a000)
[ 1704.877856] 9e40:                                     80a1cc3c 00000000 00000010 b88123c8
[ 1704.877865] 9e60: b7e69e84 80a1cc34 fff9fee9 ffffffff b7e68000 00000009 b7e69ea4 b7e69e88
[ 1704.877874] 9e80: 800704b0 80630ea4 fff9fee9 60080013 80a1cc28 fff9fee9 b7e69edc b7e69ea8
[ 1704.877884] 9ea0: 8040f558 80070460 fff9fee9 ffffffff 00000000 00000000 00000009 80a1cb7c
[ 1704.877893] 9ec0: 00000000 80a1cb7c 00000000 00000010 b7e69ef4 b7e69ee0 803e1ba4 8040f514
[ 1704.877902] 9ee0: 00000e48 80a1cb7c b7e69f14 b7e69ef8 803e1c9c 803e1b74 b88123c0 b92acb18
[ 1704.877911] 9f00: b8812790 b8d815d8 b7e69f24 b7e69f18 803e2250 803e1bc8 b7e69f5c b7e69f28
[ 1704.877921] 9f20: 80167bac 803e222c 00000000 00000000 b7e69f54 b8ab9ffc 00000000 8098c794
[ 1704.877930] 9f40: b8ab9b00 8000efc4 b7e68000 00000000 b7e69f6c b7e69f60 80167d6c 80167b28
[ 1704.877939] 9f60: b7e69f8c b7e69f70 80047d38 80167d60 b7e68000 b7e68010 8000efc4 b7e69fb0
[ 1704.877949] 9f80: b7e69fac b7e69f90 80012820 80047c84 01155490 011549a8 00000001 00000006
[ 1704.877957] 9fa0: 00000000 b7e69fb0 8000ee5c 80012790 00000000 353d8c0f 7efc4308 00000000
[ 1704.877966] 9fc0: 01155490 011549a8 00000001 00000006 00000000 00000000 76cf3ba0 00000003
[ 1704.877975] 9fe0: 00000000 7efc42e4 0002272f 76e2ed66 60080030 00000003 00000000 00000000
[ 1704.877998] [<80630ee8>] (__down_interruptible) from [<800704b0>] (down_interruptible+0x5c/0x68)
[ 1704.878015] [<800704b0>] (down_interruptible) from [<8040f558>] (vchiu_queue_push+0x50/0xd8)
[ 1704.878032] [<8040f558>] (vchiu_queue_push) from [<803e1ba4>] (send_worker_msg+0x3c/0x54)
[ 1704.878045] [<803e1ba4>] (send_worker_msg) from [<803e1c9c>] (vc_cma_set_reserve+0xe0/0x1c4)
[ 1704.878057] [<803e1c9c>] (vc_cma_set_reserve) from [<803e2250>] (vc_cma_release+0x30/0x38)
[ 1704.878069] [<803e2250>] (vc_cma_release) from [<80167bac>] (__fput+0x90/0x1e0)
[ 1704.878082] [<80167bac>] (__fput) from [<80167d6c>] (____fput+0x18/0x1c)
[ 1704.878094] [<80167d6c>] (____fput) from [<80047d38>] (task_work_run+0xc0/0xf8)
[ 1704.878109] [<80047d38>] (task_work_run) from [<80012820>] (do_work_pending+0x9c/0xc4)
[ 1704.878123] [<80012820>] (do_work_pending) from [<8000ee5c>] (work_pending+0xc/0x20)
[ 1704.878133] Code: e50b1034 e3a01000 e50b2030 e580300c (e5823000)

..the fix is to ensure that we have actually initialized the queue before we attempt
to push any items onto it.  This occurs if we do an open() followed by a close() without
any activity in between.

Signed-off-by: Colin Ian King <colin.king@canonical.com>

vchiq_arm: Sort out the vmalloc case

See: raspberrypi#1055

vchiq: hack: Add include depecated dma include file
@popcornmix Tim Gover vcsm: VideoCore shared memory service for BCM2835
Add experimental support for the VideoCore shared memory service.
This allows user processes to allocate memory from VideoCore's
GPU relocatable heap and mmap the buffers. Additionally, the memory
handles can passed to other VideoCore services such as MMAL, OpenMax
and DispmanX

* This driver was originally released for BCM28155 which has a different
  cache architecture to BCM2835. Consequently, in this release only
  uncached mappings are supported. However, there's no fundamental
  reason which cached mappings cannot be support or BCM2835
* More refactoring is required to remove the typedefs.
* Re-enable the some of the commented out debug-fs statistics which were
  disabled when migrating code from proc-fs.
* There's a lot of code to support sharing of VCSM in order to support
  Android. This could probably done more cleanly or perhaps just

Signed-off-by: Tim Gover <timgover@gmail.com>

config: Disable VC_SM for now to fix hang with cutdown kernel

vcsm: Use boolean as it cannot be built as module

On building the bcm_vc_sm as a module we get the following error:

v7_dma_flush_range and do_munmap are undefined in vc-sm.ko.

Fix by making it not an option to build as module

vcsm: Add ioctl for custom cache flushing

vc-sm: Move headers out of arch directory

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@popcornmix popcornmix vc_mem: Add vc_mem driver
Signed-off-by: popcornmix <popcornmix@gmail.com>

BCM270x: Move vc_mem

Make the vc_mem module available for ARCH_BCM2835 by moving it.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@Wren6991 @popcornmix Wren6991 Add /dev/gpiomem device for rootless user GPIO access
Signed-off-by: Luke Wren <luke@raspberrypi.org>

bcm2835-gpiomem: Fix for ARCH_BCM2835 builds

Build on ARCH_BCM2835, and fail to probe if no IO resource.

See: raspberrypi#1154
@Wren6991 @popcornmix Wren6991 Add SMI driver
Signed-off-by: Luke Wren <wren6991@gmail.com>
@Wren6991 @popcornmix Wren6991 Add SMI NAND driver
Signed-off-by: Luke Wren <wren6991@gmail.com>
@popcornmix Aron Szabo lirc: added support for RaspberryPi GPIO
lirc_rpi: Use read_current_timer to determine transmitter delay. Thanks to jjmz and others
See: raspberrypi#525

lirc: Remove restriction on gpio pins that can be used with lirc

Compute Module, for example could use different pins

lirc_rpi: Add parameter to specify input pin pull

Depending on the connected IR circuitry it might be desirable to change the
gpios internal pull from it pull-down default behaviour. Add a module
parameter to allow the user to set it explicitly.

Signed-off-by: Julian Scheel <julian@jusst.de>

lirc-rpi: Use the higher-level irq control functions

This module used to access the irq_chip methods of the
gpio controller directly, rather than going through the
standard enable_irq/irq_set_irq_type functions. This
caused problems on pinctrl-bcm2835 which only implements
the irq_enable/disable methods and not irq_unmask/mask.

lirc-rpi: Correct the interrupt usage

1) Correct the use of enable_irq (i.e. don't call it so often)
2) Correct the shutdown sequence.
3) Avoid a bcm2708_gpio driver quirk by setting the irq flags earlier

lirc-rpi: use getnstimeofday instead of read_current_timer

read_current_timer isn't guaranteed to return values in
microseconds, and indeed it doesn't on a Pi2.

Issue: linux#827

lirc-rpi: Add device tree support, and a suitable overlay

The overlay supports DT parameters that match the old module
parameters, except that gpio_in_pull should be set using the
strings "up", "down" or "off".

lirc-rpi: Also support pinctrl-bcm2835 in non-DT mode
@popcornmix popcornmix Add cpufreq driver
Signed-off-by: popcornmix <popcornmix@gmail.com>
@popcornmix popcornmix Added hwmon/thermal driver for reporting core temperature. Thanks Dorian
BCM270x: Move thermal sensor to Device Tree

Add Device Tree support to bcm2835-thermal driver.
Add thermal sensor device to Device Tree.
Don't add platform device when booting in DT mode.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@popcornmix popcornmix Add Chris Boot's i2c driver
i2c-bcm2708: fixed baudrate

Fixed issue where the wrong CDIV value was set for baudrates below 3815 Hz (for 250MHz bus clock).
In that case the computed CDIV value was more than 0xffff. However the CDIV register width is only 16 bits.
This resulted in incorrect setting of CDIV and higher baudrate than intended.
Example: 3500Hz -> CDIV=0x11704 -> CDIV(16bit)=0x1704 -> 42430Hz
After correction: 3500Hz -> CDIV=0x11704 -> CDIV(16bit)=0xffff -> 3815Hz
The correct baudrate is shown in the log after the cdiv > 0xffff correction.

Perform I2C combined transactions when possible

Perform I2C combined transactions whenever possible, within the
restrictions of the Broadcomm Serial Controller.

Disable DONE interrupt during TA poll

Prevent interrupt from being triggered if poll is missed and transfer
starts and finishes.

i2c: Make combined transactions optional and disabled by default

i2c: bcm2708: add device tree support

Add DT support to driver and add to .dtsi file.
Setup pins in .dts file.
i2c is disabled by default.

Signed-off-by: Noralf Tronnes <notro@tronnes.org>

bcm2708: don't register i2c controllers when using DT

The devices for the i2c controllers are in the Device Tree.
Only register devices when not using DT.

Signed-off-by: Noralf Tronnes <notro@tronnes.org>

I2C: Only register the I2C device for the current board revision

i2c_bcm2708: Fix clock reference counting

Fix grabbing lock from atomic context in i2c driver

2 main changes:
- check for timeouts in the bcm2708_bsc_setup function as indicated by this comment:
      /* poll for transfer start bit (should only take 1-20 polls) */
  This implies that the setup function can now fail so account for this everywhere it's called
- Removed the clk_get_rate call from inside the setup function as it locks a mutex and that's not ok since we call it from under a spin lock.

i2c-bcm2708: When using DT, leave the GPIO setup to pinctrl

i2c-bcm2708: Increase timeouts to allow larger transfers

Use the timeout value provided by the I2C_TIMEOUT ioctl when waiting
for completion. The default timeout is 1 second.

See: raspberrypi#260

i2c-bcm2708/BCM270X_DT: Add support for I2C2

The third I2C bus (I2C2) is normally reserved for HDMI use. Careless
use of this bus can break an attached display - use with caution.

It is recommended to disable accesses by VideoCore by setting
hdmi_ignore_edid=1 or hdmi_edid_file=1 in config.txt.

The interface is disabled by default - enable using the
i2c2_iknowwhatimdoing DT parameter.

bcm2708-spi: Don't use static pin configuration with DT

Also remove superfluous error checking - the SPI framework ensures the
validity of the chip_select value.

i2c-bcm2708: Remove non-DT support

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@popcornmix Vincent Sanders bcm2835: add v4l2 camera device
- Supports raw YUV capture, preview, JPEG and H264.
- Uses videobuf2 for data transfer, using dma_buf.
- Uses 3.6.10 timestamping
- Camera power based on use
- Uses immutable input mode on video encoder

Signed-off-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: Luke Diamand <luked@broadcom.com>

V4L2: Fixes from 6by9

V4L2: Fix EV values. Add manual shutter speed control

V4L2 EV values should be in units of 1/1000. Corrected.
Add support for V4L2_CID_EXPOSURE_ABSOLUTE which should
give manual shutter control. Requires manual exposure mode
to be selected first.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Correct JPEG Q-factor range

Should be 1-100, not 0-100

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Fix issue of driver jamming if STREAMON failed.

Fix issue where the driver was left in a partially enabled
state if STREAMON failed, and would then reject many IOCTLs
as it thought it was streaming.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Fix ISO controls.

Driver was passing the index to the GPU, and not the desired
ISO value.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Add flicker avoidance controls

Add support for V4L2_CID_POWER_LINE_FREQUENCY to set flicker
avoidance frequencies.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Add support for frame rate control.

Add support for frame rate (or time per frame as V4L2
inverts it) control via s_parm.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Improve G_FBUF handling so we pass conformance

Return some sane numbers for get framebuffer so that
we pass conformance.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Fix information advertised through g_vidfmt

Width and height were being stored based on incorrect

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Add support for inline H264 headers

to control H264 inline headers.
Requires firmware fix to work correctly, otherwise format
has to be set to H264 before this parameter is set.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Fix JPEG timestamp issue

JPEG images were coming through from the GPU with timestamp
of 0. Detect this and give current system time instead
of some invalid value.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Fix issue when switching down JPEG resolution.

JPEG buffer size calculation is based on input resolution.
Input resolution was being configured after output port
format. Caused failures if switching from one JPEG resolution
to a smaller one.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Enable MJPEG encoding

Requires GPU firmware update to support MJPEG encoder.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Correct flag settings for compressed formats

Set flags field correctly on enum_fmt_vid_cap for compressed
image formats.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: H264 profile & level ctrls, FPS control and auto exp pri

Several control handling updates.
H264 profile and level controls.
Timeperframe/FPS reworked to add V4L2_CID_EXPOSURE_AUTO_PRIORITY to
select whether AE is allowed to override the framerate specified.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Correct BGR24 to RGB24 in format table

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Add additional pixel formats. Correct colourspace

Adds the other flavours of YUYV, and NV12.
Corrects the overlay advertised colourspace.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Drop logging msg from info to debug

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Initial pass at scene modes.

Only supports exposure mode and metering modes.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Add manual white balance control.

Adds support for V4L2_CID_RED_BALANCE and
V4L2_CID_BLUE_BALANCE. Only has an effect if

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

config: Enable V4L / MMAL driver

V4L2: Increase the MMAL timeout to 3sec

MJPEG codec flush is now taking longer and results
in a kernel panic if the driver has stopped waiting for
the result when it finally completes.
Increase the timeout value from 1 to 3secs.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Add support for setting H264_I_PERIOD

Adds support for the parameter V4L2_CID_MPEG_VIDEO_H264_I_PERIOD
to set the frequency with which I frames are produced.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Enable GPU function for removing padding from images.

GPU can now support arbitrary strides, although may require
additional processing to achieve it. Enable this feature
so that the images delivered are the size requested.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Add support for V4L2_PIX_FMT_BGR32

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Set the colourspace to avoid odd YUV-RGB conversions

Removes the amiguity from the conversion routines and stops
them dropping back to the SD vs HD choice of coeffs.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Make video/still threshold a run-time param

Move the define for at what resolution the driver
switches from a video mode capture to a stills mode
capture to module parameters.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Fix incorrect pool sizing

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Add option to disable enum_framesizes.

Gstreamer's handling of a driver that advertises
V4L2_FRMSIZE_TYPE_STEPWISE to define the supported
resolutions is broken. See bug

Optional parameter of gst_v4l2src_is_broken added.
If non-zero, the driver claims not to support that
ioctl, and gstreamer should be happy again (it
guesses a set of defaults for itself).

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Add support for more image formats

Adds YVU420 (YV12), YVU420SP (NV21), and BGR888.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

V4L2: Extend range for V4L2_CID_MPEG_VIDEO_H264_I_PERIOD

Request to extend the range from the fairly arbitrary
1000 frames (33 seconds at 30fps). Extend out to the
max range supported (int32 value).
Also allow 0, which is handled by the codec as only
send an I-frame on the first frame and never again.
There may be an exception if it detects a significant
scene change, but there's no easy way around that.

Signed-off-by: Dave Stevenson <dsteve@broadcom.com>

bcm2835-camera: stop_streaming now has a void return

BCM2835-V4L2: Fix compliance test failures

VIDIOC_TRY_FMT and VIDIOC_S_FMT tests were faling due
to reporting V4L2_COLORSPACE_JPEG when the colour
format wasn't V4L2_PIX_FMT_JPEG.
Now reports V4L2_COLORSPACE_SMPTE170M for YUV formats.

bcm2835 camera planar/packed stride length

Added a field to the mmal_fmt struct used to compute the bytes per line
when using a particular format. This results in the correct stride being
calculated even when the format is planar.

Signed-off-by: Garrett Wilson <g@floft.net>

bcm2835: camera: check for scene not being found

static analysis by cppcheck detected some potential NULL pointer
dereference issues:

[drivers/media/platform/bcm2835/controls.c:854]: (error) Possible null
  pointer dereference: scene
  (and lines 858, 859 too)

it is possible that scene is not found because of an invalue ctrl->val
and is therefore NULL and hence causing a null pointer dereference.

Signed-off-by: Colin Ian King <colin.king@canonical.com>

bcm2835: memcpy port data to m rather than rmsg

static analysis by cppcheck detected a memcpy to rmsg which is
not actually initialized at that point.  The memcpy should be copying
to variable m instead.

Signed-off-by: Colin Ian King <colin.king@canonical.com>

BCM2835-V4L2: Return buffers to videobuf2 on shutdown

Fixes the kernel warning from videobuf2 as buffers
are now returned as they are being flushed on

squash: Fixup bcm2835-camera for changes in kernel 4.4 api
@notro @popcornmix notro firmware: bcm2835: Support ARCH_BCM270x
Support booting without Device Tree.
Turn on USB power.
Load driver early because of lacking support for deferred probing
in many drivers.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@notro @popcornmix notro char: broadcom: Add vcio module
Add module for accessing the mailbox property channel through
/dev/vcio. Was previously in bcm2708-vcio.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@pelwell @popcornmix pelwell fdt: Add support for the CONFIG_CMDLINE_EXTEND option 24365c8
@pelwell @popcornmix pelwell scripts: Add mkknlimg and knlinfo scripts from tools repo
The Raspberry Pi firmware looks for a trailer on the kernel image to
determine whether it was compiled with Device Tree support enabled.
If the firmware finds a kernel without this trailer, or which has a
trailer indicating that it isn't DT-capable, it disables DT support
and reverts to using ATAGs.

The mkknlimg utility adds that trailer, having first analysed the
image to look for signs of DT support and the kernel version string.

knlinfo displays the contents of the trailer in the given kernel image.

scripts/mkknlimg: Add support for ARCH_BCM2835

Add a new trailer field indicating whether this is an ARCH_BCM2835
build, as opposed to MACH_BCM2708/9. If the loader finds this flag
is set it changes the default base dtb file name from bcm270x...
to bcm283y...

Also update knlinfo to show the status of the field.

scripts/mkknlimg: Improve ARCH_BCM2835 detection

The board support code contains sufficient strings to be able to
distinguish 2708 vs. 2835 builds, so remove the check for
bcm2835-pm-wdt which could exist in either.

Also, since the canned configuration is no longer built in (it's
a module), remove the config string checking.

See: raspberrypi#1157
@notro @popcornmix notro BCM2708: Add core Device Tree support
Add the bare minimum needed to boot BCM2708 from a Device Tree.

Signed-off-by: Noralf Tronnes <notro@tronnes.org>

BCM2708: DT: change 'axi' nodename to 'soc'

Change DT node named 'axi' to 'soc' so it matches ARCH_BCM2835.
The VC4 bootloader fills in certain properties in the 'axi' subtree,
but since this is part of an upstreaming effort, the name is changed.

Signed-off-by: Noralf Tronnes notro@tronnes.org

BCM2708_DT: Correct length of the peripheral space

Use dts-dirs feature for overlays.

The kernel makefiles have a dts-dirs target that is for vendor subdirectories.

Using this fixes the install_dtbs target, which previously did not install the overlays.

BCM270X_DT: configure I2S DMA channels

Signed-off-by: Matthias Reichl <hias@horus.com>

BCM270X_DT: switch to bcm2835-i2s

I2S soundcard drivers with proper devicetree support (i.e. not linking
to the cpu_dai/platform via name but to cpu/platform via of_node)
will work out of the box without any modifications.

When the kernel is compiled without devicetree support the platform
code will instantiate the bcm2708-i2s driver and I2S soundcard drivers
will link to it via name, as before.

Signed-off-by: Matthias Reichl <hias@horus.com>

SDIO-overlay: add poll_once-boolean parameter

Add paramter to toggle sdio-device-polling
done every second or once at boot-time.

Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>

BCM270X_DT: Make mmc overlay compatible with current firmware

The original DT overlay logic followed a merge-then-patch procedure,
i.e. parameters are applied to the loaded overlay before the overlay
is merged into the base DTB. This sequence has been changed to
patch-then-merge, in order to support parameterised node names, and
to protect against bad overlays. As a result, overrides (parameters)
must only target labels in the overlay, but the overlay can obviously target nodes in the base DTB.

mmc-overlay.dts (that switches back to the original mmc sdcard
driver) is the only overlay violating that rule, and this patch
fixes it.

bcm270x_dt: Use the sdhost MMC controller by default

The "mmc" overlay reverts to using the other controller.

squash: Add cprman to dt

BCM270X_DT: Use clk_core for I2C interfaces
@notro @popcornmix notro bcm2835: Match with BCM2708 Device Trees
Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@popcornmix popcornmix Add non-mainline source for rtl8192cu wireless driver version v4.0.2_…
…9000 as this is widely used. Disabled older rtlwifi driver

8192cu needs old wireless extensions

The obsolete WIRELESS_EXT configuration is used
by the old Realtek code and is needed for AP support.

8192cu: CONFIG_AP_MODE hardcoded in autoconf.h
@ssvb @popcornmix ssvb fbdev: add FBIOCOPYAREA ioctl
Based on the patch authored by Ali Gholami Rudi at

Provide an ioctl for userspace applications, but only if this operation
is hardware accelerated (otherwide it does not make any sense).

Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
@popcornmix popcornmix rtl8192cu: Add PID for D-Link DWA 131 f3dfa06
@marckleinebudde @popcornmix marckleinebudde rtl8192c_rf6052: PHY_RFShadowRefresh(): fix off-by-one
Signed-off-by: Marc Kleine-Budde <mkl@blackshift.org>
@popcornmix popcornmix Allow mac address to be set in smsc95xx
Signed-off-by: popcornmix <popcornmix@gmail.com>
@hglm @popcornmix hglm Speed up console framebuffer imageblit function
Especially on platforms with a slower CPU but a relatively high
framebuffer fill bandwidth, like current ARM devices, the existing
console monochrome imageblit function used to draw console text is
suboptimal for common pixel depths such as 16bpp and 32bpp. The existing
code is quite general and can deal with several pixel depths. By creating
special case functions for 16bpp and 32bpp, by far the most common pixel
formats used on modern systems, a significant speed-up is attained
which can be readily felt on ARM-based devices like the Raspberry Pi
and the Allwinner platform, but should help any platform using the
fb layer.

The special case functions allow constant folding, eliminating a number
of instructions including divide operations, and allow the use of an
unrolled loop, eliminating instructions with a variable shift size,
reducing source memory access instructions, and eliminating excessive
branching. These unrolled loops also allow much better code optimization
by the C compiler. The code that selects which optimized variant is used
is also simplified, eliminating integer divide instructions.

The speed-up, measured by timing 'cat file.txt' in the console, varies
between 40% and 70%, when testing on the Raspberry Pi and Allwinner
ARM-based platforms, depending on font size and the pixel depth, with
the greater benefit for 32bpp.

Signed-off-by: Harm Hanemaaijer <fgenfb@yahoo.com>
@popcornmix popcornmix enabling the realtime clock 1-wire chip DS1307 and 1-wire on GPIO4 (a…
…s a module)

1-wire: Add support for configuring pin for w1-gpio kernel module
See: raspberrypi#457

Add bitbanging pullups, use them for w1-gpio

Allows parasite power to work, uses module option pullup=1

bcm2708: Ensure 1-wire pullup is disabled by default, and expose as module parameter

Signed-off-by: Alex J Lennon <ajlennon@dynamicdevices.co.uk>

w1-gpio: Add gpiopin module parameter and correctly free up gpio pull-up pin, if set

Signed-off-by: Alex J Lennon <ajlennon@dynamicdevices.co.uk>

w1-gpio: Sort out the pullup/parasitic power tangle
@popcornmix popcornmix Added Device IDs for August DVB-T 205 f3954b0
@popcornmix popcornmix config: Enable CONFIG_MEMCG, but leave it disabled (due to memory cos…
…t). Enable with cgroup_enable=memory.
@koalo @popcornmix koalo ASoC: Add support for PCM5102A codec
Some definitions to support the PCM5102A codec
by Texas Instruments.

Signed-off-by: Florian Meier <florian.meier@koalo.de>
@koalo @popcornmix koalo ASoC: Add support for HifiBerry DAC
This adds a machine driver for the HifiBerry DAC.
It is a sound card that can
be stacked onto the Raspberry Pi.

Signed-off-by: Florian Meier <florian.meier@koalo.de>
@koalo @popcornmix koalo ASoC: Add support for Rpi-DAC bc372df
@popcornmix Daniel Matuschek ASoC: wm8804: Implement MCLK configuration options, add 32bit support…
… WM8804 can run with PLL frequencies of 256xfs and 128xfs for most sample rates. At 192kHz only 128xfs is supported. The existing driver selects 128xfs automatically for some lower samples rates. By using an additional mclk_div divider, it is now possible to control the behaviour. This allows using 256xfs PLL frequency on all sample rates up to 96kHz. It should allow lower jitter and better signal quality. The behavior has to be controlled by the sound card driver, because some sample frequency share the same setting. e.g. 192kHz and 96kHz use 24.576MHz master clock. The only difference is the MCLK divider.

This also added support for 32bit data.

Signed-off-by: Daniel Matuschek <daniel@matuschek.net>
@popcornmix Daniel Matuschek ASoC: BCM:Add support for HiFiBerry Digi. Driver is based on the patc…
…hed WM8804 driver.

Signed-off-by: Daniel Matuschek <daniel@matuschek.net>

Add a parameter to turn off SPDIF output if no audio is playing

This patch adds the paramater auto_shutdown_output to the kernel module.
Default behaviour of the module is the same, but when auto_shutdown_output
is set to 1, the SPDIF oputput will shutdown if no stream is playing.

bugfix for 32kHz sample rate, was missing

HiFiBerry Digi: set SPDIF status bits for sample rate

The HiFiBerry Digi driver did not signal the sample rate in the SPDIF status bits.
While this is optional, some DACs and receivers do not accept this signal. This patch
adds the sample rate bits in the SPDIF status block.
@popcornmix Daniel Matuschek ASoC: wm8804: Set idle_bias_off to false Idle bias has been change to…
… remove warning on driver startup

Signed-off-by: Daniel Matuschek <daniel@matuschek.net>
@iqaudio @popcornmix iqaudio Add IQaudIO Sound Card support for Raspberry Pi
Set a limit of 0dB on Digital Volume Control

The main volume control in the PCM512x DAC has a range up to
+24dB. This is dangerously loud and can potentially cause massive
clipping in the output stages. Therefore this sets a sensible
limit of 0dB for this control.
@popcornmix popcornmix hid: Reduce default mouse polling interval to 60Hz
Reduces overhead when using X
@popcornmix Daniel Matuschek Added support for HiFiBerry DAC+
The driver is based on the HiFiBerry DAC driver. However HiFiBerry DAC+ uses
a different codec chip (PCM5122), therefore a new driver is necessary.
@popcornmix Daniel Matuschek Added driver for HiFiBerry Amp amplifier add-on board
The driver contains a low-level hardware driver for the TAS5713 and the
drivers for the Raspberry Pi I2S subsystem.

TAS5713: return error if initialisation fails

Existing TAS5713 driver logs errors during initialisation, but does not return
an error code. Therefore even if initialisation fails, the driver will still be
loaded, but won't work. This patch fixes this. I2C communication error will now
reported correctly by a non-zero return code.

HiFiBerry Amp: fix device-tree problems

Some code to load the driver based on device-tree-overlays was missing. This is added by this patch.
@nismoryco @popcornmix nismoryco Update ds1307 driver for device-tree support
Signed-off-by: Ryan Coe <bluemrp9@gmail.com>
@pelwell @popcornmix pelwell BCM270x_DT: Add pwr_led, and the required "input" trigger
The "input" trigger makes the associated GPIO an input.  This is to support
the Raspberry Pi PWR LED, which is driven by external hardware in normal use.

N.B. pwr_led is not available on Model A or B boards.

leds-gpio: Implement the brightness_get method

The power LED uses some clever logic that means it is driven
by a voltage measuring circuit when configured as input, otherwise
it is driven by the GPIO output value. This patch wires up the
brightness_get method for leds-gpio so that user-space can monitor
the LED value via /sys/class/gpio/led1/brightness. Using the input
trigger this returns an indication of the system power health,
otherwise it is just whatever value the trigger has written most

See: raspberrypi#1064
@pelwell @popcornmix pelwell enc28j60: Add device tree compatible string and an overlay e2346dc
@popcornmix Waldemar Brodkorb Add driver for rpi-proto
Forward port of 3.10.x driver from https://github.com/koalo
We are using a custom board and would like to use rpi 3.18.x
kernel. Patch works fine for our embedded system.

URL to the audio chip:

Playback tested with devicetree enabled.

Signed-off-by: Waldemar Brodkorb <wbrodkorb@conet.de>
@popcornmix popcornmix config: Add default configs eb2a7b9
@notro @popcornmix notro bcm2835: bcm2835_defconfig
Some options in bcm2835_defconfig are now the default and
some have changed. Update to keep functionality.

No longer available: SCSI_MULTI_LUN and RESOURCE_COUNTERS.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

bcm2835: bcm2835_defconfig enable MMC_BCM2835

Enable the downstream bcm2835-mmc driver and DMA support.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

bcm2835: bcm2835_defconfig enable BCM2708_MBOX

Enable the mailbox driver.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

bcm2835: bcm2835_defconfig use FB_BCM2708

Enable the bcm2708 framebuffer driver.
Disable the simple framebuffer driver, which matches the
device handed over by u-boot.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

bcm2835: Merge bcm2835_defconfig with bcmrpi_defconfig

These commands where used to make this commit:

./scripts/diffconfig -m arch/arm/configs/bcm2835_defconfig arch/arm/configs/bcmrpi_defconfig > merge.cfg

cat << EOF > filter

grep -F -v -f filter merge.cfg > filtered.cfg

cat << EOF > added.cfg

ARCH=arm scripts/kconfig/merge_config.sh arch/arm/configs/bcm2835_defconfig filtered.cfg added.cfg
ARCH=arm make oldconfig

ARCH=arm make savedefconfig
cp defconfig arch/arm/configs/bcm2835_defconfig

rm merge.cfg filter filtered.cfg added.cfg defconfig

ARCH=arm make bcm2835_defconfig
ARCH=arm make oldconfig

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

configs: Incorporate v4.1 dependency changes

Commit 78e9b7d added a
dependency on TI_ST instead of selecting it, disabling:

Commit 652ccae added a
depency on ARM_CRYPTO, disabling:

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>


bcm2835: Sync bcm2835_defconfig with bcmrpi_defconfig

These commands where used to make this commit:

: Get changed and new config values from a merge
./scripts/diffconfig -m arch/arm/configs/bcm2835_defconfig arch/arm/configs/bcmrpi_defconfig > merge.cfg

: Remove these options
cat << EOF > filter

: Apply filter
grep -F -v -f filter merge.cfg > filtered.cfg

: Add these options
: watchdog contains the restart/poweroff code.
cat << EOF > added.cfg

: Create new config
ARCH=arm scripts/kconfig/merge_config.sh arch/arm/configs/bcm2835_defconfig filtered.cfg added.cfg
: Verify
ARCH=arm make oldconfig

: Update bcm2835_defconfig
ARCH=arm make savedefconfig
cp defconfig arch/arm/configs/bcm2835_defconfig

: Clean up
rm merge.cfg filter filtered.cfg added.cfg defconfig

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@pelwell @popcornmix pelwell Improve __copy_to_user and __copy_from_user performance
Provide a __copy_from_user that uses memcpy. On BCM2708, use
optimised memcpy/memmove/memcmp/memset implementations.

arch/arm: Add mmiocpy/set aliases for memcpy/set

See: raspberrypi#1082
@ghollingworth @popcornmix ghollingworth rpi-ft5406: Add touchscreen driver for pi LCD display
Fix driver detection failure Check that the buffer response is non-zero meaning the touchscreen was detected

rpi-ft5406: Use firmware API
@pelwell @popcornmix pelwell gpio-poweroff: Allow it to work on Raspberry Pi
The Raspberry Pi firmware manages the power-down and reboot
process. To do this it installs a pm_power_off handler, causing
the gpio-poweroff module to abort the probe function.

This patch introduces a "force" DT property that overrides that
behaviour, and also adds a DT overlay to enable and control it.

Note that running in an active-low configuration (DT parameter
"active_low") requires a custom dt-blob.bin and probably won't
allow a reboot without switching off, so an external inversion
of the trigger signal may be preferable.
@pelwell @popcornmix pelwell spidev: Add "spidev" compatible string to silence warning a105ea5
@pelwell @popcornmix pelwell Merge pull request #1043 from XECDesign/sense-4.0
mfd: Add Raspberry Pi Sense HAT core driver
@pelwell @popcornmix pelwell scripts/dtc: Add overlay support 90dbf1c
@popcornmix Jan Grulich tpa6130a2: Add headphone switch control
Signed-off-by: Jan Grulich <jan@grulich.eu>
@popcornmix Jan Grulich RaspiDAC3 support
Signed-off-by: Jan Grulich <jan@grulich.eu>

config: fix RaspiDAC Rev.3x dependencies

Change depends to SND_BCM2708_SOC_I2S || SND_BCM2835_SOC_I2S
like the other I2S soundcard drivers.

Signed-off-by: Matthias Reichl <hias@horus.com>
@popcornmix popcornmix irq-bcm2835: Fix building with 2708 bf2ddba
@popcornmix P33M rpi_display: add backlight driver and overlay
Add a mailbox-driven backlight controller for the Raspberry Pi DSI
touchscreen display. Requires updated GPU firmware to recognise the
mailbox request.

Signed-off-by: Gordon Hollingworth <gordon@raspberrypi.org>
@pelwell @popcornmix pelwell scripts: Multi-platform support for mkknlimg and knlinfo
The firmware uses tags in the kernel trailer to choose which dtb file
to load. Current firmware loads bcm2835-*.dtb if the '283x' tag is true,
otherwise it loads bcm270*.dtb. This scheme breaks if an image supports
multiple platforms.

This patch adds '270X' and '283X' tags to indicate support for RPi and
upstream platforms, respectively. '283x' (note lower case 'x') is left
for old firmware, and is only set if the image only supports upstream
@HiassofT @popcornmix HiassofT bcm2835-dma: Fix up convert to DMA pool aadadcb
@anholt @popcornmix anholt drm/vc4: Add suport for 3D rendering using the V3D engine.
This is a squash of the out-of-tree development series.  Since that
series contained code from the first "get a demo triangle rendered
using a hacked up driver using binary shader code" to "plug the last
known security hole", it's hard to reconstruct a different series of
incremental development that's mergeable without security holes
throughout it.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Force HDMI to connected.
For some reason on the downstream tree, the HPD GPIO isn't working.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: bo cache locking fixes.
Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: bo cache locking cleanup.
Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Use job_lock to protect seqno_cb_list.
We're (mostly) not supposed to be using struct_mutex in drivers these

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Drop struct_mutex around CL validation.
We were using it so that we could make sure that shader validation
state didn't change while we were validating, but now shader
validation state is immutable.  The bcl/rcl generation doesn't do any
other BO dereferencing, and seems to have no other global state
dependency not covered by job_lock / bo_lock.

Fixes a lock order reversal between mmap_sem and struct_mutex.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Drop struct_mutex around CL validation.
We were using it so that we could make sure that shader validation
state didn't change while we were validating, but now shader
validation state is immutable.  The bcl/rcl generation doesn't do any
other BO dereferencing, and seems to have no other global state
dependency not covered by job_lock / bo_lock.  We only need to hold
struct_mutex for object unreferencing.

Fixes a lock order reversal between mmap_sem and struct_mutex.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Add support for more display plane formats.
Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: No need to stop the stopped threads.
This was leftover debug code from the hackdriver.  We never submit
unless the thread is already idle.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Remove extra barrier()s aroudn CTnCA/CTnEA setup.
The writel() that these expand to already does barriers.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Fix a typo in a V3D debug register.
Signed-off-by: Eric Anholt <eric@anholt.net>
@pelwell @popcornmix pelwell drm/vc4: Enable VC4 modules, and increase CMA size with overlay
If using the overlay, be careful not to boot to GUI or run startx,
or the Pi will almost hang, reporting stalls in kernel threads.
@popcornmix popcornmix squash: fixups 9021edc
@popcornmix popcornmix squash: add missing vc4-kms-v3d-overlay.dtb to makefile 21f3ea3
@anholt @popcornmix anholt clk: bcm2835: Also build the driver for downstream kernels.
Signed-off-by: Eric Anholt <eric@anholt.net>
@hsteinhaus @popcornmix hsteinhaus dts: Added overlay for gpio_ir_recv driver ac35dc3
@mwilliams03 @popcornmix mwilliams03 New overlay for PiScreen2r 188695a
@ali1234 @popcornmix ali1234 Build i2c_gpio module and add a device tree overlay to configure it. bdbd75a
@popcornmix Stuart MacLean Add support for the HiFiBerry DAC+ Pro.
The HiFiBerry DAC+ and DAC+ Pro products both use the existing bcm sound driver with the DAC+ Pro having a special clock device driver representing the two high precision oscillators.

An addition bug fix is included for the PCM512x codec where by the physical size of the sample frame is used in the calculation of the LRCK divisor as it was found to be wrong when using 24-bit depth sample contained in a little endian 4-byte sample frame.
@pelwell @popcornmix pelwell BCM270X_DT: Add at86rf233 overlay
Add an overlay to support the Atmel AT86RF233 WPAN transceiver on spi0.0.

See: raspberrypi#1151
@ondrej1024 @popcornmix ondrej1024 dts: Added overlay for Adafruit PiTFT 2.8" capacitive touch screen 7c2080d
@anholt @popcornmix anholt drm: Put an optional field in the driver struct for GEM obj struct size.
This allows a driver to derive from the CMA object without copying all
of the code.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt mm: Remove the PFN busy warning
See commit dae803e -- the warning is
expected sometimes when using CMA.  However, that commit still spams
my kernel log with these warnings.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Add an interface for capturing the GPU state after a hang.
This can be parsed with vc4-gpu-tools tools for trying to figure out
what was going on.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Update a bunch of code to match upstream submission.
This gets almost everything matching, except for the MSAA support and
using generic PM domains.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Add support for MSAA rendering.
For MSAA, you set a bit in the binner that halves the size of tiles in
each direction, so you can pack 4 samples per pixel in the tile
buffer.  During rendering, you can load and store raw tile buffer
contents (to save the per-sample MSAA contents), or you can load/store
resolved tile buffer contents (loads spam the pixel value to all 4
samples, and stores either average the 4 color samples, or store the
first sample for Z/S).

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: A few more non-functional changes to sync to upstream.
At this point all that's left is the force-enable of HDMI connector,
and using direct firmware calls to turn on V3D instead of the generic
power domain support.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm: Use the driver's gem_object_free function from CMA helpers.
VC4 wraps the CMA objects in its own structures, so it needs to do its
own teardown (waiting for GPU to finish, updating bo_stats tracking).
The other CMA drivers are using drm_gem_cma_free_object as their
gem_free_object, so this should be a no-op for them.

Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Synchronize validation code for v2 submission upstream.
Signed-off-by: Eric Anholt <eric@anholt.net>
@anholt @popcornmix anholt drm/vc4: Use "hpd-gpios" for HDMI GPIO, like what landed upstream.
Signed-off-by: Eric Anholt <eric@anholt.net>
pelwell commented Dec 28, 2015

I have no objections to the Pi Zero dtb. I think the inverted activity LED, although different, is a sensible use of a single LED.

mobluse commented Dec 29, 2015

I think this should be tested and working on Pi Model A before it is merged, because maybe there are many A:s that can use this if it was adapted to these.


I increased "g-np-tx-fifo-size" because 16 seemed "a bit small" compared to the legacy code, but from inspection of the code it seems to remain unused. My personal preference is to make "g-rx-fifo-size" as large as possible, and set "g-tx-fifo-size" based on actual testing.

notro commented Dec 29, 2015

Regarding OTG on model A

Have anyone tried setting dr_mode = "peripheral"; in the Device Tree?
Maybe it's enough, I don't know. If it doesn't work then we need to know which register holds the value of the USB_OTGID pin and see how it's used in the driver (AFAICT this pin is tied low on both A and B).


static int dwc2_core_reset(struct dwc2_hsotg *hsotg)
        if (hsotg->dr_mode == USB_DR_MODE_HOST) {
                gusbcfg = readl(hsotg->regs + GUSBCFG);
                gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
                gusbcfg |= GUSBCFG_FORCEHOSTMODE;
                writel(gusbcfg, hsotg->regs + GUSBCFG);
        } else if (hsotg->dr_mode == USB_DR_MODE_PERIPHERAL) {
                gusbcfg = readl(hsotg->regs + GUSBCFG);
                gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
                gusbcfg |= GUSBCFG_FORCEDEVMODE;
                writel(gusbcfg, hsotg->regs + GUSBCFG);
        } else if (hsotg->dr_mode == USB_DR_MODE_OTG) {
                gusbcfg = readl(hsotg->regs + GUSBCFG);
                gusbcfg &= ~GUSBCFG_FORCEHOSTMODE;
                gusbcfg &= ~GUSBCFG_FORCEDEVMODE;
                writel(gusbcfg, hsotg->regs + GUSBCFG);

How should a type A to type A cable be wired to test this? Straight through?

lurch commented Dec 29, 2015

How should a type A to type A cable be wired to test this? Straight through?

I'm still away from home so can't check, but IIRC when I was experimenting with rpiboot on a ModelA, I used a cable like this and carefully removed the outer shielding and snipped through the 5V wire.

@JanLuca @popcornmix JanLuca MMC: Do not use mmc_debug if CONFIG_MMC_BCM2835 is not set
If CONFIG_MMC_BCM2835 was not set the compiling of the kernel failed
since mmc_debug was not defined but used in drivers/mmc/core/quirks.c.

This patch add a ifdef-check for CONFIG_MMC_BCM2835 to the change of
commit 64d3954
mobluse commented Dec 29, 2015

About Pi A: I tested with an adapter similar to this: http://www.amazon.com/niceEshop-Changer-Adapter-Coupler-Converter/dp/B00BPHBCX6 . It has pin 1 connected to pin 1, 2 to 2, 3 to 3, 4 to 4, shield to shield (one contact is up side down compared to the other). It gives this with dmesg in Ubuntu 14.04 LTS and similar in Raspbian:
[382925.640098] usb 6-1: new full-speed USB device number 4 using uhci_hcd
[382925.806151] usb 6-1: New USB device found, idVendor=0a5c, idProduct=2763
[382925.806163] usb 6-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[382925.806171] usb 6-1: Product: BCM2708 Boot
[382925.806178] usb 6-1: Manufacturer: Broadcom

I also looked at the schematics for Pi A and B on page 3 and USB_OTGID/K16 is connected to ground: https://www.raspberrypi.org/documentation/hardware/raspberrypi/schematics/Raspberry-Pi-Rev-2.1-Model-AB-Schematics.pdf

gbaman commented Dec 29, 2015

Using a similar adapter to @mobluse, just gave it a go with my A+. Unfortunately not having any luck getting it to appear on attached computer. Have tried this pull request and also tried using a kernel with dwc2 set to Gadget only mode. Both without any luck on fresh Raspbian images.

notro commented Dec 30, 2015

Device mode on A+

I haven't made that cable yet and I don't know when I get the time, but I believe I have found a solution to overriding the USB_OTGID pin. So if someone want to test it, here's what I intend to try:

  * Returns the mode of operation, host or device
 static inline int dwc2_is_host_mode(struct dwc2_hsotg *hsotg)
+   if (hsotg->dr_mode == USB_DR_MODE_HOST)
+       return 1;
+   else if (hsotg->dr_mode == USB_DR_MODE_PERIPHERAL)
+       return 0;
    return (readl(hsotg->regs + GINTSTS) & GINTSTS_CURMODE_HOST) != 0;
 static inline int dwc2_is_device_mode(struct dwc2_hsotg *hsotg)
-   return (readl(hsotg->regs + GINTSTS) & GINTSTS_CURMODE_HOST) == 0;
+   return !dwc2_is_host_mode(hsotg)

This is also needed in DT: dr_mode = "peripheral";

The GINTSTS reg is available in debugfs, host mode is determined by bit 0:

# keyboard connected on zero
~$ sudo grep GINTSTS /sys/kernel/debug/20980000.usb/regdump
GINTSTS = 0x04000021

# keyboard disconnected
~$ sudo grep GINTSTS /sys/kernel/debug/20980000.usb/regdump
GINTSTS = 0x04008022

@notro can you rebase and drop the LED commit for now.
Feel free to submit it as separate PR if you are still keen, but no point holding up this PR for it.

notro commented Jan 1, 2016

and drop the LED commit for now.

Yes there was a strong opposition against that, and I do see the advantage as a combined poweron/disk activity indicator. And it's possible to change the polarity with a dt parameter so I have changed my position on this.

@ED6E0F17 mentioned a patch usb: dwc2: gadget: fix TX FIFO size and address initialization which I had a look at but it wasn't a shallow rabbithole so I dropped that. Some notes I made: https://gist.github.com/notro/1abf4b2d89580d80489a

I'll see if I can rebase later today. I have made a dwc_otg overlay for the Zero, but I'm wondering if I should do a dwc2 overlay as well which can be used as a debug aid with usb problems to rule out dwc_otg as the problem. It can also have parameters to set the fifo properties.

notro added some commits Dec 26, 2015
@notro notro bcmrpi_defconfig: Enable the dwc2 driver and some gadget drivers.
This enables us to use the USB On-the-Go support on Pi Zero.

Other changes:
CONFIG_BT_HCIUART is needed because of this commit: 7841d06
CONFIG_HW_RANDOM_BCM2835=y is not needed since it's the default.
I was unable to track down why IIO_BUFFER_CB changed from builtin to a module.
This is what I get when I try to change it to builtin:
  This feature depends on another which has been configured as a module.
  As a result, this feature will be built as a module.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
@notro notro bcm2708_dt: Add a Pi Zero Device Tree file
This is identical to bcm2708-rpi-b-plus.dts except for:
* dwc2 is used for usb instead of dwc_otg to get gadget support
* Power led node is removed

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>

I could imagine a dwc2 overlay being useful on a Pi2, for debugging USB issues.
That driver may improve with time (perhaps getting upstream fiq support) so may be a more viable option in the future.


@P33M can you comment on the fifo sizes? Is it something we can (and should) be setting, or is it just read only?

notro commented Jan 1, 2016

Version 2

  • Go back to inverted act led
  • Add dwc_otg overlay so the Pi Zero can use the dwc_otg driver in host mode with it's fiq support
  • Add dwc2 overlay with parameters for fifo sizes

I won't pursue Device mode on A+. I'm returning to Pi2 and ARCH_BCM2835 for another go.

P33M commented Jan 2, 2016

According to the databook, the register for device periodic per-endpoint FIFO size is a fixed quantity with a variable start address. The fixed quantity is read-only and in the top 16 bits of DPTXFSIZn (offset 0x104 + (EPn-1) * 4). The power-on reset value for the bottom bits is a bit complicated and I'm not sure I can post without disclosing internal core details.

However, these bottom bits are writeable which means you can change the start address for each of these endpoint FIFOs as long as the value written doesn't exceed the value loaded on reset.

As far as I'm concerned, treating the periodic FIFO sizes in device mode as read-only (incl. POR value which implies a contiguous block of SRAM) is the right thing to do.

@clivem clivem referenced this pull request in raspberrypi/firmware Jan 3, 2016

dtparam=act_led_activelow=off should be inverted for a Zero #524


@notro we'd like to get this in. Can you rebase?
I wonder whether defaulting to dwc2 on Pi Zero is correct. Currently it's not an issue as you need to choose the pi zero dtb explicitly, but it may make sense to have firmware choose that automatically which would then mean Pi Zero users who don't care about gadget support will get dwc2 driver which may not work so well. Any thoughts?

Anyway, I'd like to get this in soon and a BRANCH=next firmware build out for testing.

clivem commented Jan 8, 2016

No, please do not default to dwc2! ;) People who want that need to specify it via config.txt, IMHO.

pelwell commented Jan 8, 2016


Currently it's not an issue as you need to choose the pi zero dtb explicitly

clivem commented Jan 8, 2016

Phil.... LOL. I am relaxed! Or I was until I read the @popcornmix comment, "may make sense to have firmware choose that automatically which would then mean Pi Zero users who don't care about gadget support will get dwc2 driver which may not work so well." And he's right, about the not working so well part. Trying to use a mt7601u wireless dongle was a complete train-wreck for me with dwc2. ;)

pelwell commented Jan 8, 2016

We wouldn't make that switch unless we thought dwc2 was solid. An interim step is likely to be to load the zero dtb automatically, but the defer the dwc2 selection to an overlay.

clivem commented Jan 8, 2016

@pelwell ...... which makes perfect sense to me! Any thought of dwc2 by default, was the scary part! ;)

pelwell commented Jan 8, 2016

We don't hate you that much.

notro commented Jan 8, 2016

@notro we'd like to get this in. Can you rebase?

I have very little time to work on the Pi these days and I have a kernel build in my tree that is not trivial to stash away.
Can you use the patches you already applied to rpi-4.4.y and just drop the zero dts patch?

I wonder whether defaulting to dwc2 on Pi Zero is correct.

I assumed dwc2 was stable since it had moved out of staging, but @clivem has proven otherwise at least for one use case. Yes it's probably best to make it optional, also not having fiq support generates a lot of "slow" interrupts.

This means that the only difference between bcm2708-rpi-b-plus.dts and bcm2708-rpi-zero.dts is the pwr_led node. So maybe we should just drop the zero dts for now.
I wonder if this would work:

#include "bcm2708-rpi-b-plus.dts"

/ {
    model = "Raspberry Pi Model Zero";

/ {
    __overrides__ {
        /delete-property/ pwr_led_gpio;
        /delete-property/ pwr_led_activelow;
        /delete-property/ pwr_led_trigger;

/delete-node/ &pwr_led;


sudo BRANCH=next rpi-update
should get the latest kernel with these commits.

@popcornmix popcornmix added a commit to raspberrypi/firmware that referenced this pull request Jan 9, 2016
@popcornmix popcornmix kernel: Bump to 4.4.0-rc8
kernel: Add dwc2 overlay for Pi Zero
See: raspberrypi/linux#1239
@popcornmix popcornmix added a commit to Hexxeh/rpi-firmware that referenced this pull request Jan 9, 2016
@popcornmix popcornmix kernel: Bump to 4.4.0-rc8
kernel: Add dwc2 overlay for Pi Zero
See: raspberrypi/linux#1239
NicoHood commented Jan 9, 2016

Ahm just a small question:
I tried the command above (with a clean full jessie image) and this installs kernel 4.4.0-rc8+ . I then set the device tree in config.txt and loaded the g_ether module via /etc/modules.

#config.txt (appended, the rest was kept unchanged)

Its not working the same way as @gbaman s tutorial did. So something must be missing, possibly the modules?

clivem commented Jan 9, 2016

Erm..... I haven't actually looked at that next branch, but I think your config.txt should be more like ....

and "dtoverlay=mmc" if you'd prefer to revert back to mmc rather than sdhost driver.

..... if you want to use g_ether on Zero.

You might want to wait for @popcornmix to reply before you blindly follow my advice and risk ending up with something that doesn't boot. ;)

NicoHood commented Jan 9, 2016 edited

Okay so an even smaller tutorial now, but with more details after installing added.

Thanks for all the help! I just wrote down all my notes, most work was done by other great people!

Also an still unanswered question is, if the pi zero is protected if I power it from the OTG port and the nativ power port at the same time. Otherwise switching from host to device at runtime (with separate PSU) is dangerous and might kill one of the devices connected to it.

Edit: Here is a solution: #1212 (comment)

nmaas87 commented Jan 9, 2016

Hey @NicoHood

regarding the question about the MAC address,
it should work like this (change in /etc/network/interfaces)

allow-hotplug usb0
iface usb0 inet static
hwaddress ether 00:01:04:1b:2C:1F

According to this site: http://www.youritronics.com/how-to-set-the-mac-address-from-etcnetworkinterfaces-in-debian/ (have not tested it, yet)

Regarding the powering problem, I think the only way to be sure would be to see, wheter there are some kind of protection diodes sepearting both (potential) power sources or such. Problem is, the electrical schematics of the pi zero are not (yet) available online (at least not where the rest of the stuff is found at: https://www.raspberrypi.org/documentation/hardware/raspberrypi/schematics/README.md )

NicoHood commented Jan 9, 2016

You can do

echo "options g_ether host_addr=00:11:22:33:44:55" | sudo tee /etc/modprobe.d/g_ether.conf

sudo nano /etc/modprobe.d/g_ether.conf
options g_ether host_addr=00:11:22:33:44:55
options g_ether dev_addr=00:66:77:88:99:00

Which works perfectly for me. I dont know which one is more reliable. the "host" setting sets the mac, that the connected host PC sees. The "dev" setting (just replace host with dev) sets the mac that the pi sees itself for its own usb0 adapter. This is confusing me. Shouldnt they be the same anyways? You can modify them in the way that the host sees the pi with a different mac than the pi sees itself.

Too bad about the schematics. Someone who inspected the pcb told me to better be carefully since it looks to him that the lines are just 1:1 connected. Thats a pity.

nmaas87 commented Jan 9, 2016

Thanks for the heads up, regarding the MAC address 👍 :)!

I think the host_addr and dev_addr are important in the fact that they are different. Yes, normally you are completly right, a ethernet/network interface is only allowed to have one mac address. But, I think the OTG Ethernet does emulate a "complete network" in that sense that you can see the OTG connection between host pc and pi zero not only as ONE network interface, but instead two.
So you got the ethernet interface on your pizero, which has the dev_addr and you got another ethernet interface on the pc, which needs to have a different mac, the host_addr. And those two virtual ethernet interfaces do communicate with each other.

HOST PC ------------------------------------------- OTG Ethernet ----------------------------------------------- PiZero
HOST PC --- HOST Virtual Ethernet Card --- Virtual Wire ---- PiZero Virtual Ethernet Card ---- PiZero

So that would make sense - but I could be complelty wrong here.

About the OTG Power - yes, that would make sense - regarding the little space they got on the pcb as well as the price point (yes, not really expensive to do stuff like this - but every cm² and cent/penny counts!). So - I would really refrain from testing OTG + USB Power on a PiZero. Chances are the poor bugger would just die (maybe not instantly, but in given time :/)

NicoHood commented Jan 9, 2016

You are totally right, that MAC explanation makes sense. So I think we got a nice module solution here. I like it more than the network interface setting.

I just have no idea why the module doesnt load at startup. I could provide additional information if you like to. This new feature looks very promising and runs very smooth so far.

nmaas87 commented Jan 9, 2016

Ok, I tried your small tutorial and can reproduce your problem:
While I could do the update without problems, the reboot after the kernel update already "broke" my setup ;). After that, I inserted

# This is required in the config.txt (make sure to use "zero" now)
# Regarding the readme this is applied automatically for a pi zero (and yes it works without)
# dtoverlay=dwc2
# The new kernel causes sd card errors for me, so i use the old sd card driver
# dtoverlay=mmc

into my /boot/config.txt and tried to boot again: Nothing. Or well - at least something: My windows machine at least did mention that there was an usb device which could not be recognized, which proofed your idea that the module is not loaded correctly via the /etc/modules file (yes, I got the g_ether entry there - of course - as well :)). Last try was to add modprobe g_ether to my /etc/rc.local - and - voilá - it came up - not a problem :). So you're right - I don't know what it is - but the entry in /etc/modules is somehow "ignored" (technically, I am sure it is not ignored, but some dependency is not yet loaded at that point of the boot process or something else happens which causes the g_ether to be not loaded if done via /etc/modules...). So I think maybe we could get some more infos via dmesg and other log files. Problem: I did neither got my Serial TTL device, nor my mini usb cable with me - so debugging that matter while beeing in need of the virtual ethernet will be a little bit hard :/.

Ah, I got the problem. Just a moment, I figure out what is really needed, but I got it working again via "only /etc/modules" :)

nmaas87 commented Jan 9, 2016

OK, solution for the "g_ether does not work, even though included in /etc/modules" is to include following lines in in /etc/modules:


dwc2 is needed and not loaded automatically(!). So the basic problem is: modprobe tries to include g_ether, but the dependency of it (dwc2) is missing and that fails. If you're trying to do that via sudo modprobe (probably later in the boot process) - dwc2 was already loaded by some other stuff - at least that is me - guessing what is going on.

@NicoHood Could you test that? Would that also work for your setup?

shrx commented Jan 9, 2016

Why do we need the full Jessie image instead of lite?

nmaas87 commented Jan 9, 2016

@shrx I think the lite Jessie image should work too - basically the needed stuff is included within the kernel - so if you upgrade that, you should have it all together :). I used the latest version of the stripped down Minibian Jessie and it worked ( https://minibianpi.wordpress.com/ ).

notro commented Jan 9, 2016

dwc2 is needed and not loaded automatically

It is, but later in the boot process. When /etc/modules is parsed it is not loaded yet, so since g_ether depends on it, it has to be manually loaded.

I believe udev Coldplug all Devices is the one autoloading dwc2, but this service is started after Load Kernel Modules and takes longer to finish:

[    7.603262] systemd[1]: Starting Load Kernel Modules...
         Starting Load Kernel Modules...
[    7.635960] systemd[1]: Starting udev Coldplug all Devices...
         Starting udev Coldplug all Devices...
[  OK  ] Started Load Kernel Modules.
[    8.060231] systemd[1]: Started Load Kernel Modules.
[  OK  ] Started udev Coldplug all Devices.
[    8.339931] systemd[1]: Started udev Coldplug all Devices.
         Starting /etc/rc.local Compatibility...
[  OK  ] Started /etc/rc.local Compatibility.

/etc/modules is hooked up to systemd through a symlink:

~$ ls -l /etc/modules-load.d/modules.conf
lrwxrwxrwx 1 root root 10 Sep  5 22:23 /etc/modules-load.d/modules.conf -> ../modules

Another thing:
Now I remember why I wanted dwc2 to be the default on Zero, because coupled with a raspian systemd service that would only run only on a zero in usb device mode, it could load g_serial automatically and give an out-of-the-box serial console: Just connect the Zero to a computer and get a console. But all in due time...

NicoHood commented Jan 9, 2016

I updated the tutorial above now with a lot more details!

It works great, and with the patch @nmaas87 suggested everything works now at boot. Plug in your pi and it always connects with the same mac now.

@shrx I guess we dont. I just tried with the full install. Feel free to test the lite image. I guess it should work without any problems.

I guess its quite easy to use OTG now ;)
The power issue is still now solved though.
I also never tried to switch back to a non-zero pi. Not sure what will happen then with this kinda dev state.

clivem commented Jan 9, 2016

@notro LOL. You're making me nostalgic! I remember how cool I thought it was when I plugged my first BeagleBone into a PC and got a serial console....... But the fact doesn't change, without dwc2 receiving some love, while it may be the answer for world peace in device mode, it leaves something to be desired, (that's the politest way of putting it), in host mode. wifi dongles suddenly disconnecting from USB bus is the tip of an iceberg. Heaven help you, (train crash leading to toxic chemical spill), trying to use it for USB audio! ;)

notro commented Jan 9, 2016

I remember how cool I thought it was when I plugged my first BeagleBone into a PC and got a serial console

Me too, it was so easy to get started.

Your comment jogged my Solution-Finder-Unit and I realised that my goal might not be so far away after all.
config.txt is easily available for editing, so a user could add: dtoverlay=dwc2
Then the service could do:

if zero() and dwc2() and device_mode() then
    modprobe g_serial

A keyboard plugged in after boot will still work, and if this behaviour is not desirable, the service can just be disabled.

So here's project for me when I'm done banging my head agaist getting Pi2 mainline support working without issues (unless someone can't wait that long and finds a way to actually implement this).

Heaven help you, (train crash leading to toxic chemical spill), trying to use it for USB audio! ;)

My focus on an easily available serial console gave me tunnel vision, so I'm glad there's a review process :-)

NicoHood commented Jan 9, 2016

Just a small question:
I cannot use ping without sudo on the g_ether module. Is this normal?

pi@raspberrypizero:~$ ping
ping: icmp open socket: Operation not permitted
nmaas87 commented Jan 9, 2016

@NicoHood Not a problem because of g_ether I think. Test if the SUID bit is set correctly on your ping command, as described here: http://ben.goodacre.name/tech/Ping:_icmp_open_socket:_Operation_not_permitted_(Linux)
On my Minibian Jessie - it worked right out of the box :)
Edit: It did work right out of the box (ping without root access, however, the SUID bit was not set on my PiZero (I have no idea how this could work, because network access is always a kind of root thing...). On my Ubuntu Box, however, the SUID bit is set on /bin/ping :).


I now also tried the usb gadget. It seems to work fine, I tried to install an OS into the emulated usb drive, but it failed after booting. But it was used on a slow "random" sd card, I will try this again on a better one.

The question is now, how to enable the gadgets by default on a zero, but disable them in a pi2? We currently have to edit the config.txt. Is there a way to add something like if inside those files, or can the be done with a kernel fix?

Also can we automatically enable the dwc2 module if g_ether is loaded for example? Would this work too? Then people do not have to type both of them in the /etc/modules.

I really would like to see this feature more flexible across different pi models. This is really awesome. And some of the responsible pi people should solve the power issue by providing a schematic, so we know if connecting a PSU in device mode hurts any device.


I just tried again (with a different sd card, from the beginning) and after applying:

echo "device_tree=bcm2708-rpi-zero.dtb" | sudo tee -a /boot/config.txt

The raspi stops booting and the rainbow screen only shows up. If i remove the line (with an sd card reader) it boots up again.

Using g_ether without those line does not work, but the line itself now also doesnt work anymore. So what has changed?

I first setup the device with a pi2, however the 4.4 kernel all boots fine on a zero, just the device tree setting makes it not boot up anymore.

I put in my old sd card and I noticed that bcm2708-rpi-zero.dtb is just missing on the boot partition. Maybe because I did the rpi-upgrade on the pi2, not the zero? I guess this is a bug?

t3chguy commented Jan 15, 2016

@NicoHood no, the zero specific .dtb is no more as I found out. Just dtoverlay=dwc2 is enough after sudo BRANCH=next rpi-update, then either manually load one of the g_ modules or stick


inside /etc/modules

@NicoHood NicoHood referenced this pull request in Hexxeh/rpi-firmware Jan 15, 2016
@popcornmix popcornmix kernel: Bump to 4.4.0
firmware: Fix for h264 picture corruption with german tv
See: raspberrypi/firmware#499

firmware: TC358762: Avoid hanging when no DISPLAY_I2C_PORT defined in blob
See: raspberrypi/firmware#526

firmware: video_render: Ensure pixel_x/pixel_y don't overflow calculations
See: raspberrypi/firmware#525

firmware: arm_loader: Set uart0_clkrate before merging overlays
See: https://www.raspberrypi.org/forums/viewtopic.php?f=29&t=113753&sid=df96fc384c4794d3344e5462a2ab0c45&start=150#p879523
notro commented Jan 16, 2016

What's the link to the Raspian github issue tracker? I can't find it.
I have a proposal on how to easily enable a usb console.

lurch commented Jan 16, 2016

What's the link to the Raspian github issue tracker?

I guess https://github.com/RPi-Distro/repo/issues is the nearest match?


Correct. For raspbian build of default debian packages it is:

For raspberry pi specific packages and init scripts (not from debian) then:

notro commented Jan 16, 2016

@t3chguy thanks! I updated the tutorial above. echo "dtoverlay=dwc2" | sudo tee -a /boot/config.txt is now all you need.

Is there a chance to load this overlay by default or is there any disadvantage for this?

There is also another bug that /etc/modprobe.d/g_ether.conf is not recognized sometimes. I think its mostly after first loading g_ether. If you remove it and load it again it will be applied.

lurch commented Jan 16, 2016

Is there a chance to load this overlay by default or is there any disadvantage for this?

See the messages from @clivem earlier in this discussion.


So its unstable for host mode on some devices and it will be kept like this?

Cant we load/unload those overlays at runtime? (Sorry i never did such kernel things before).
Is there any chance to see this somehow improved in the future or is this just the limitation we have?

@notro notro closed this Jan 27, 2016

Hey, what about using with usbip client as host on Raspberry Pi Zero (as usbip server running in another rasbperry pi with a device connected), then wirelessly you can plug Raspberry Pi Zero on a computer and see the device if it where connected directly to the computer. Can this be done?

lurch commented Mar 15, 2016

@braian87b I've read your comment several times, and I'm still not quite sure what you're asking for?

I think what you're suggesting is having a PiZero in gadget-mode connected over USB to a PC, and then having the PiZero communicating wirelessly with another Raspberry Pi, and then having a USB device you connect to this second Raspberry Pi "tunnelled" wirelessly to the PiZero, and then exposing it as a 'virtual' USB device to the PC which the PiZero is connected to?

However the PiZero only has one USB OTG port, which can operate in either host or device mode. So if you've got the Pi connected in device-mode to a PC, there's then no 'spare' USB host port into which to plug a Wifi dongle. There are ways of communicating with a wifi chip over the Pi's GPIO pins e.g. https://shop.pimoroni.com/products/esp8266-phat but that only runs at 1mbps, whereas USB2 runs at up to 480mbps (even USB1 is 12mbps).


@lurch Right... sorry, I didn't notice... thanks!

dtaht commented Jun 9, 2016

I was catching up on this discussion in light of some discussion elsewhere ( https://plus.google.com/u/0/+EricRaymond/posts/DR2zNNyNs3Z )

and I see my pi2 had in 4.4.12+/

for dwc2.ko

but not in my latest update (4.4.12-v7). Does the OTG stuff work on the pi2 or pi3?

lurch commented Jun 9, 2016

Does the OTG stuff work on the pi2 or pi3?

No. The USB hub in the LAN9514 'gets in the way' and means that the single USB port in the SoC on those devices can only work in host (master) mode. It's only because the PiZero has a single USB port (i.e. no USB hub) that it's also able to (optionally) work in device (slave) mode.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment