- 03 May, 2019 40 commits
-
-
Thinh Nguyen authored
The setting of the dr_mode may need to check the controller's revision. The revision is set in the dwc3_core_is_valid(), which comes after dr_mode setting. Let's move it closer to the start of the dwc3_probe() function and before calling dwc3_get_dr_mode(). Signed-off-by:
Thinh Nguyen <thinhn@synopsys.com> Signed-off-by:
Felipe Balbi <felipe.balbi@linux.intel.com>
-
Thinh Nguyen authored
All DWC3 controllers are LPM capable. Report that in the usb_gadget.lpm_capable for the gadget driver to properly output the bcdUSB value in the descriptor. Signed-off-by:
-
Jonas Bonn authored
If the USB device is connected to a host, the CPU cannot be suspended or else the USB device appears to be disconnected from the host's point of view. Only after a "USB suspend" state has been entered (as set by the host) or the host is disconnected can the system safely be suspended: in both these states, the clock is stopped. As such, this patch associates a "wake lock" with the running clock of the UDC to keep the system awake as long as the host maintains the USB connection active. Signed-off-by:
Jonas Bonn <jonas@norrbonn.se> CC: Cristian Birsan <cristian.birsan@microchip.com> CC: Felipe Balbi <balbi@kernel.org> CC: Greg Kroah-Hartman <gregkh@linuxfoundation.org> CC: Nicolas Ferre <nicolas.ferre@microchip.com> CC: Alexandre Belloni <alexandre.belloni@bootlin.com> CC: Ludovic Desroches <ludovic.desroches@microchip.com> CC: linux-arm-kernel@lists.infradead.org CC: linux-usb@vger.kernel.org Signed-off-by:
-
Jonas Bonn authored
This patch adds support for USB suspend to the Atmel UDC. When suspended, the UDC clock can be stopped, resulting in some power savings. The "wake up" interrupt will fire irregardless of whether the clock is running or not, allowing the UDC clock to be restarted when the USB master wants to wake the device again. The IRQ state of this device is somewhat fiddly. The "wake up" IRQ seems to actually be a "bus activity" indicator; the IRQ is almost continuously asserted so enabling this IRQ should only be done after a suspend when the wake IRQ becomes relevant. Similarly, the "suspend" IRQ detects "bus inactivity" and may therefore fire together with a "wake" if the two types of activity coincide during the period between two IRQ handler invocations; therefore, it's important to ignore the "suspend" IRQ while waiting for a wake-up. This has been tested on a SAMA5D2 board. Signed-off-by:
-
Jonas Bonn authored
This patch adds set and clear functions for enabling/disabling interrupts. This simplifies the implementation a bit as the masking of previously set bits doesn't need to be so explicit. Signed-off-by:
-
Minas Harutyunyan authored
Fix calculation of transfer size on completion in function dwc2_gadget_get_xfersize_ddma(). Added increment of descriptor pointer to move to next descriptor in the loop. Fixes: aa3e8bc8 ("usb: dwc2: gadget: DDMA transfer start and complete") Signed-off-by:
Minas Harutyunyan <hminas@synopsys.com> Signed-off-by:
-
Minas Harutyunyan authored
If core not supported lpm, i.e. BCM2835 then confusing warnings seen in log. To avoid these warnings, added function dwc2_set_param_lpm() to set lpm and other lpm related parameters based on lpm support by core. Signed-off-by:
-
Minas Harutyunyan authored
Channel disabling/halting should performed for enabled only channels to avoid warnings "Unable to clear enable on channel N" which seen if host works in Slave mode. Signed-off-by:
-
Minas Harutyunyan authored
On ISOC transfer completion, in DDMA mode, set actual frame number returning to function driver in usb_request. Due to core limitation, returning frame number is 11-bit wide. Signed-off-by:
-
Sergey Senozhatsky authored
A trivial patch. cpu_to_le16() is capable enough to detect __builtin_constant_p() and to use an appropriate compile time ___constant_swahbXX() function. So we can use cpu_to_le16() instead of __constant_cpu_to_le16(). Signed-off-by:
Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Signed-off-by:
-
Minas Harutyunyan authored
Some function drivers queueing more than 128 ISOC requests at a time. To avoid "descriptor chain full" cases, increasing descriptors count from MAX_DMA_DESC_NUM_GENERIC to MAX_DMA_DESC_NUM_HS_ISOC for ISOC's only. Signed-off-by:
-
Chunfeng Yun authored
In some places, the code prints a human-readable USB endpoint transfer type (e.g. "bulk"). This involves a switch statement sometimes wrapped around in ({ ... }) block leading to code repetition. To make this scenario easier, here introduces usb_ep_type_string() function, which returns a human-readable name of provided endpoint type. It also changes a few places switch was used to use this new function. Signed-off-by:Chunfeng Yun <chunfeng.yun@mediatek.com> Signed-off-by:
-
Marek Szyprowski authored
dwc3_gadget_suspend() is called under dwc->lock spinlock. In such context calling synchronize_irq() is not allowed. Move the problematic call out of the protected block to fix the following kernel BUG during system suspend: BUG: sleeping function called from invalid context at kernel/irq/manage.c:112 in_atomic(): 1, irqs_disabled(): 128, pid: 1601, name: rtcwake 6 locks held by rtcwake/1601: #0: f70ac2a2 (sb_writers#7){.+.+}, at: vfs_write+0x130/0x16c #1: b5fe1270 (&of->mutex){+.+.}, at: kernfs_fop_write+0xc0/0x1e4 #2: 7e597705 (kn->count#60){.+.+}, at: kernfs_fop_write+0xc8/0x1e4 #3: 8b3527d0 (system_transition_mutex){+.+.}, at: pm_suspend+0xc4/0xc04 #4: fc7f1c42 (&dev->mutex){....}, at: __device_suspend+0xd8/0x74c #5: 4b36507e (&(&dwc->lock)->rlock){....}, at: dwc3_gadget_suspend+0x24/0x3c irq event stamp: 11252 hardirqs last enabled at (11251): [<c09c54a4>] _raw_spin_unlock_irqrestore+0x6c/0x74 hardirqs last disabled at (11252): [<c09c4d44>] _raw_spin_lock_irqsave+0x1c/0x5c softirqs last enabled at (9744): [<c0102564>] __do_softirq+0x3a4/0x66c softirqs last disabled at (9737): [<c0128528>] irq_exit+0x140/0x168 Preemption disabled at: [<00000000>] (null) CPU: 7 PID: 1601 Comm: rtcwake Not tainted 5.0.0-rc3-next-20190122-00039-ga3f4ee4f8a52 #5252 Hardware name: SAMSUNG EXYNOS (Flattened Device Tree) [<c01110f0>] (unwind_backtrace) from [<c010d120>] (show_stack+0x10/0x14) [<c010d120>] (show_stack) from [<c09a4d04>] (dump_stack+0x90/0xc8) [<c09a4d04>] (dump_stack) from [<c014c700>] (___might_sleep+0x22c/0x2c8) [<c014c700>] (___might_sleep) from [<c0189d68>] (synchronize_irq+0x28/0x84) [<c0189d68>] (synchronize_irq) from [<c05cbbf8>] (dwc3_gadget_suspend+0x34/0x3c) [<c05cbbf8>] (dwc3_gadget_suspend) from [<c05bd020>] (dwc3_suspend_common+0x154/0x410) [<c05bd020>] (dwc3_suspend_common) from [<c05bd34c>] (dwc3_suspend+0x14/0x2c) [<c05bd34c>] (dwc3_suspend) from [<c051c730>] (platform_pm_suspend+0x2c/0x54) [<c051c730>] (platform_pm_suspend) from [<c05285d4>] (dpm_run_callback+0xa4/0x3dc) [<c05285d4>] (dpm_run_callback) from [<c0528a40>] (__device_suspend+0x134/0x74c) [<c0528a40>] (__device_suspend) from [<c052c508>] (dpm_suspend+0x174/0x588) [<c052c508>] (dpm_suspend) from [<c0182134>] (suspend_devices_and_enter+0xc0/0xe74) [<c0182134>] (suspend_devices_and_enter) from [<c0183658>] (pm_suspend+0x770/0xc04) [<c0183658>] (pm_suspend) from [<c0180ddc>] (state_store+0x6c/0xcc) [<c0180ddc>] (state_store) from [<c09a9a70>] (kobj_attr_store+0x14/0x20) [<c09a9a70>] (kobj_attr_store) from [<c02d6800>] (sysfs_kf_write+0x4c/0x50) [<c02d6800>] (sysfs_kf_write) from [<c02d594c>] (kernfs_fop_write+0xfc/0x1e4) [<c02d594c>] (kernfs_fop_write) from [<c02593d8>] (__vfs_write+0x2c/0x160) [<c02593d8>] (__vfs_write) from [<c0259694>] (vfs_write+0xa4/0x16c) [<c0259694>] (vfs_write) from [<c0259870>] (ksys_write+0x40/0x8c) [<c0259870>] (ksys_write) from [<c0101000>] (ret_fast_syscall+0x0/0x28) Exception stack(0xed55ffa8 to 0xed55fff0) ... Fixes: 01c10880 ("usb: dwc3: gadget: synchronize_irq dwc irq in suspend") Signed-off-by:Marek Szyprowski <m.szyprowski@samsung.com> Signed-off-by:
-
Andy Shevchenko authored
When we read an array of integers from device properties, the temporary buffer is allocated. However, in case of dwc3_set_incr_burst_type() it's not freed. Free allocated buffer immediately after use. Signed-off-by:
Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by:
-
Robin Murphy authored
Now that the bulk API has a "get all of the clocks" helper to match what this code wants, there's little reason not to switch over. Signed-off-by:
Robin Murphy <robin.murphy@arm.com> Signed-off-by:
-
Minas Harutyunyan authored
Added delayed status support for Control transfers. Tested in all 3 modes: Slave, BDMA and DDMA. Performed tests: USB CV (Ch9 and MSC), Control Read/Write tests using Synopsys USB test environment function driver. Signed-off-by:
-
Alexandre Belloni authored
There is no actual need to do the enable/disable_irq dance. Instead enable the interrupts on the phy only when necessary. Signed-off-by:
Alexandre Belloni <alexandre.belloni@bootlin.com> Signed-off-by:
-
Alexandre Belloni authored
The STOTG04 phy is used as a drop-in replacement of the ISP1301 but some bits doesn't have exactly the same meaning and this can lead to issues. Detect the phy dynamically and avoid writing to reserved bits. Signed-off-by:
-
Alexandre Belloni authored
Only INT_VBUS_VLD is set to generate ATX interrupts on the phy but INT_SESS_VLD is checked in vbus_work. This leads to cases where hot-plugging USB doesn't work after boot. Signed-off-by:
-
Alexandre Belloni authored
Use a threaded IRQ to handle vbus_work instead of using the global worqueue. Signed-off-by:
-
Alexandre Belloni authored
Simplify .probe and .remove by using devm managed allocations and requests. Signed-off-by:
-
Jules Maselbas authored
Most of the phy initialization is shared between host and gadget, this adds the turnaround configuration only used by gadgets to the global phy init. Acked-by:
Jules Maselbas <jmaselbas@kalray.eu> Signed-off-by:
-
Jules Maselbas authored
As the phy initialization is almost the same in host and gadget mode. This only move the phy initialization functions into core.c for now, the goal is to share theses functions between the two modes. Acked-by:
-
Jules Maselbas authored
The phy utmi width information is already set in hsotg params, phyif is only used in few places and I don't see any reason to not use hsotg's params. Moreover the utmi width was being forced to 16 bits by platform initialization which doesn't take in account HW configuration. Acked-by:
-
Jules Maselbas authored
The function dwc2_hsotg_init is only called once just before calling dwc2_hsotg_core_init_disconnected which does the same initialization: setting the usbcfg register with turnaround time, timeout calibration and phy width. Acked-by:
-
Jules Maselbas authored
Makes GHWCFG4_UTMI_PHY_DATA* defines closer to their relative shift and mask defines to improve readability. Acked-by:
-
Douglas Anderson authored
The "host" USB port on rk3288 has a hardware errata where we've got to assert a PHY reset whenever we see a remote wakeup. Add that quirk property to the device tree. Signed-off-by:
Douglas Anderson <dianders@chromium.org> Reviewed-by:
Matthias Kaehlcke <mka@chromium.org> Signed-off-by:
-
Douglas Anderson authored
Let's hook up the resets to the three USB PHYs on rk3288 as per the bindings. This is in preparation for a future patch that will set the "snps,reset-phy-on-wake" on the host port. Signed-off-by:
-
Douglas Anderson authored
On the rk3288 USB host-only port (the one that's not the OTG-enabled port) the PHY can get into a bad state when a wakeup is asserted (not just a wakeup from full system suspend but also a wakeup from autosuspend). We can get the PHY out of its bad state by asserting its "port reset", but unfortunately that seems to assert a reset onto the USB bus so it could confuse things if we don't actually deenumerate / reenumerate the device. We can also get the PHY out of its bad state by fully resetting it using the reset from the CRU (clock reset unit), which does a more full reset. The CRU-based reset appears to actually cause devices on the bus to be removed and reinserted, which fixes the problem (albeit in a hacky way). It's unfortunate that we need to do a full re-enumeration of devices at wakeup time, but this is better than alternative of letting the bus get wedged. Signed-off-by:
Yunzhi Li <lyz@rock-chips.com> Signed-off-by:
-
Douglas Anderson authored
On Rockchip rk3288 there's a hardware quirk where we need to assert the reset signal to the PHY when we get a remote wakeup on one of the two ports. Document this quirk in the bindings. Signed-off-by:
-
Martin Blumenstingl authored
Various boards have an external VBUS supply regulator. This regulator depends on the current mode of the controller which is defined as: - dr_mode set to either "host" or "peripheral" (fixed value) - dr_mode set to "otg", based on the OTG status the dwc2 controller internally switches between "host" and "peripheral" mode (selection happens at runtime) Based on the current mode the regulator has to be enabled or disabled: - host: provide power to the connected USB device, thus the regulator has to be enabled - peripheral: the host device to which the controller is connected provides power, thus the regulator has to be disabled Add the dt-bindings documentation for this property so .dts authors know that this property exists and how it behaves. Fixes: 531ef5eb ("usb: dwc2: add support for host mode external vbus supply") Acked-by:
Rob Herring <robh@kernel.org> Signed-off-by:
Martin Blumenstingl <martin.blumenstingl@googlemail.com> Signed-off-by:
-
Alan Stern authored
The net2280 UDC driver (and also net2272, probably via copy-and-paste) incorrectly checks the req->zero flag during OUT transfers, after copying data from the UDC's FIFO into memory. This makes no sense at all; the "zero" flag indicates that an extra zero-length packet should be appended to an IN transfer if the length is an even multiple of the maxpacket size. It has nothing to do with OUT transfers. In practice this doesn't cause any problems because gadget drivers never set req->zero for OUT transfers anyway. Still, it is an error and unnecessary code, so this patch removes the check. Signed-off-by:
Alan Stern <stern@rowland.harvard.edu> Signed-off-by:
-
Alan Stern authored
The net2280 driver includes an unnecessary test for an endpoint's queue being empty. The test is redundant; it sits inside a conditional block of an "if" statement which already tests the endpoint's queue. This patch removes the redundant test. Signed-off-by:
-
Arnd Bergmann authored
The dependency to ensure this driver links correctly fails since it can not be a loadable module: drivers/usb/phy/phy-fsl-usb.o: In function `fsl_otg_set_peripheral': phy-fsl-usb.c:(.text+0x2224): undefined reference to `usb_gadget_vbus_disconnect' Make the option 'tristate' so it can work correctly. Fixes: 5a8d651a ("usb: gadget: move gadget API functions to udc-core") Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
Romain Izard authored
To be able to use the default USB class drivers available in Microsoft Windows, we need to add OS descriptors to the exported USB gadget to tell the OS that we are compatible with the built-in drivers. Copy the OS descriptor support from f_rndis into f_ncm. As a result, using the WINNCM compatible ID, the UsbNcm driver is loaded on enumeration without the need for a custom driver or inf file. Signed-off-by:
Romain Izard <romain.izard.pro@gmail.com> Signed-off-by:
-
Romain Izard authored
When connecting a CDC-NCM gadget to an host that uses the NTP-32 mode, or that relies on the default CRC setting, the current implementation gets confused, and does not expect the correct signature for its packets. Fix this, by ensuring that the ndp_sign member in the f_ncm structure always contain a valid value. Signed-off-by:
-
Chunfeng Yun authored
When the driver tries to get optional clock, it ignores all errors, but if only ignores -ENOENT, it will cover some real errors, such as -EPROBE_DEFER, so use devm_clk_get_optional() to get optional clock. Cc: Minas Harutyunyan <hminas@synopsys.com> Signed-off-by:
-
Minas Harutyunyan authored
Avoiding switch to L1 state in any stage of control transfers. Send NYET handshake to LPM token. Renamed GLPMCFG_LPM_ACCEPT_CTRL_ISOC to GLPMCFG_LPM_REJECT_CTRL_CONTROL because by setting this bit core reject LPM token. Signed-off-by:
-
Douglas Anderson authored
In (e583d9db USB: global suspend and remote wakeup don't mix) we introduced wakeup_enabled_descendants() as a static function. We'd like to use this function in USB controller drivers to know if we should keep the controller on during suspend time, since doing so has a power impact. Signed-off-by:
-
Douglas Anderson authored
This is an attempt to rehash commit 0cf884e8 ("usb: dwc2: add bus suspend/resume for dwc2") on ToT. That commit was reverted in commit b0bb9bb6 ("Revert "usb: dwc2: add bus suspend/resume for dwc2"") because apparently it broke the Altera SOCFPGA. With all the changes that have happened to dwc2 in the meantime, it's possible that the Altera SOCFPGA will just magically work with this change now. ...and it would be good to get bus suspend/resume implemented. This change is a forward port of one that's been living in the Chrome OS 3.14 kernel tree. Signed-off-by:
-
