kernel.git/drivers/usb/chipidea/udc.c, branch linux-5.8.y

usb: chipidea: udc: add software sg list support

2020-05-06T01:42:35Z

The chipidea controller doesn't support short transfer for sg list, so we still keep setting IOC per TD, otherwise, there will be no interrupt for short transfer. Each TD has five entries for data buffer, each data buffer could be non-countinuous 4KB buffer, so it could handle up to 5 sg buffers one time. The benefit of this patch is avoiding OOM for low memory system(eg, 256MB) during large USB transfers, see below for detail. The non-sg handling has not changed. ufb: page allocation failure: order:4, mode:0x40cc0(GFP_KERNEL|__GFP_COMP), nodemask=(null),cpuset=/,mems_allowed=0 CPU: 2 PID: 370 Comm: ufb Not tainted 5.4.3-1.1.0+g54b3750d61fd #1 Hardware name: NXP i.MX8MNano DDR4 EVK board (DT) Call trace: dump_backtrace+0x0/0x140 show_stack+0x14/0x20 dump_stack+0xb4/0xf8 warn_alloc+0xec/0x158 __alloc_pages_slowpath+0x9cc/0x9f8 __alloc_pages_nodemask+0x21c/0x280 alloc_pages_current+0x7c/0xe8 kmalloc_order+0x1c/0x88 __kmalloc+0x25c/0x298 ffs_epfile_io.isra.0+0x20c/0x7d0 ffs_epfile_read_iter+0xa8/0x188 new_sync_read+0xe4/0x170 __vfs_read+0x2c/0x40 vfs_read+0xc8/0x1a0 ksys_read+0x68/0xf0 __arm64_sys_read+0x18/0x20 el0_svc_common.constprop.0+0x68/0x160 el0_svc_handler+0x20/0x80 el0_svc+0x8/0xc Mem-Info: active_anon:2856 inactive_anon:5269 isolated_anon:12 active_file:5238 inactive_file:18803 isolated_file:0 unevictable:0 dirty:22 writeback:416 unstable:0 slab_reclaimable:4073 slab_unreclaimable:3408 mapped:727 shmem:7393 pagetables:37 bounce:0 free:4104 free_pcp:118 free_cma:0 Node 0 active_anon:11436kB inactive_anon:21076kB active_file:20988kB inactive_file:75216kB unevictable:0kB isolated(ano Node 0 DMA32 free:16820kB min:1808kB low:2260kB high:2712kB active_anon:11436kB inactive_anon:21076kB active_file:2098B lowmem_reserve[]: 0 0 0 Node 0 DMA32: 508*4kB (UME) 242*8kB (UME) 730*16kB (UM) 21*32kB (UME) 5*64kB (UME) 2*128kB (M) 0*256kB 0*512kB 0*1024kB Node 0 hugepages_total=0 hugepages_free=0 hugepages_surp=0 hugepages_size=1048576kB Node 0 hugepages_total=0 hugepages_free=0 hugepages_surp=0 hugepages_size=32768kB Node 0 hugepages_total=0 hugepages_free=0 hugepages_surp=0 hugepages_size=2048kB Node 0 hugepages_total=0 hugepages_free=0 hugepages_surp=0 hugepages_size=64kB 31455 total pagecache pages 0 pages in swap cache Swap cache stats: add 0, delete 0, find 0/0 Free swap = 0kB Total swap = 0kB 65536 pages RAM 0 pages HighMem/MovableOnly 10766 pages reserved 0 pages cma reserved 0 pages hwpoisoned Reviewed-by: Jun Li Signed-off-by: Peter Chen

usb: chipidea: introduce CI_HDRC_CONTROLLER_VBUS_EVENT glue layer use

2020-05-06T01:41:55Z

Some vendors glue layer need to handle some events for vbus, eg, some i.mx platforms (imx7d, imx8mm, imx8mn, etc) needs vbus event to handle charger detection, its charger detection is finished at glue layer code, but not at USB PHY driver. Signed-off-by: Peter Chen

Merge 5.6-rc7 into usb-next

2020-03-23T07:04:08Z

We need the USB fixes in here as well. Signed-off-by: Greg Kroah-Hartman

usb: chipidea: udc: fix sleeping function called from invalid context

2020-03-17T19:24:28Z

The code calls pm_runtime_get_sync with irq disabled, it causes below warning: BUG: sleeping function called from invalid context at wer/runtime.c:1075 in_atomic(): 1, irqs_disabled(): 128, non_block: 0, pid: er/u8:1 CPU: 1 PID: 37 Comm: kworker/u8:1 Not tainted 20200304-00181-gbebfd2a5be98 #1588 Hardware name: NVIDIA Tegra SoC (Flattened Device Tree) Workqueue: ci_otg ci_otg_work [] (unwind_backtrace) from [] 1/0x14) [] (show_stack) from [] 5/0x94) [] (dump_stack) from [] +0xeb/0x118) [] (___might_sleep) from [] esume+0x75/0x78) [] (__pm_runtime_resume) from [] 0x23/0x74) [] (ci_udc_pullup) from [] nect+0x2b/0xcc) [] (usb_gadget_connect) from [] _connect+0x59/0x104) [] (ci_hdrc_gadget_connect) from [] ssion+0x43/0x48) [] (ci_udc_vbus_session) from [] s_connect+0x17/0x9c) [] (usb_gadget_vbus_connect) from [] bd/0x128) [] (ci_otg_work) from [] rk+0x149/0x404) [] (process_one_work) from [] 0xf7/0x3bc) [] (worker_thread) from [] x118) [] (kthread) from [] (ret_from_fork+0x11/0x34) Tested-by: Dmitry Osipenko Cc: #v5.5 Fixes: 72dc8df7920f ("usb: chipidea: udc: protect usb interrupt enable") Reported-by: Dmitry Osipenko Signed-off-by: Peter Chen Link: https://lore.kernel.org/r/20200316031034.17847-2-peter.chen@kernel.org Signed-off-by: Greg Kroah-Hartman

usb: chipidea: udc: using structure ci_hdrc device for runtime PM

2020-03-16T03:13:47Z

At current code, it doesn't maintain ci->gadget.dev's runtime PM status well, eg, during the PM operation, the PM counter for ci->gadget.dev doesn't be changed accordingly. In this commit, we use ci_hdrc device instead of ci->gadget.dev for runtime PM APIs at udc driver, in the way, we handle runtime PM APIs using unify device structure between core and udc driver. Reviewed-by: Jun Li Signed-off-by: Peter Chen

usb: chipidea: udc: protect usb interrupt enable

2019-11-18T08:45:30Z

We hit the problem with below sequence: - ci_udc_vbus_session() update vbus_active flag and ci->driver is valid, - before calling the ci_hdrc_gadget_connect(), usb_gadget_udc_stop() is called by application remove gadget driver, - ci_udc_vbus_session() will contine do ci_hdrc_gadget_connect() as gadget_ready is 1, so udc interrupt is enabled, but ci->driver is NULL. - USB connection irq generated but ci->driver is NULL. As udc irq only should be enabled when gadget driver is binded, so add spinlock to protect the usb irq enable for vbus session handling. Signed-off-by: Jun Li Signed-off-by: Peter Chen

usb: chipidea: udc: add new API ci_hdrc_gadget_connect

2019-11-18T08:45:30Z

This API is used enable device function, it is called at below situations: - VBUS is connected during boots up - Hot plug occurs during runtime Signed-off-by: Peter Chen Signed-off-by: Jun Li

usb: Spelling s/disconnet/disconnect/

2019-11-04T14:53:01Z

Fix misspellings of "disconnect". Signed-off-by: Geert Uytterhoeven Acked-by: Peter Chen Link: https://lore.kernel.org/r/20191024152747.30617-1-geert+renesas@glider.be Signed-off-by: Greg Kroah-Hartman

Merge 5.3-rc7 into usb-next

2019-09-02T17:31:18Z

We need the usb fixes in here for testing Signed-off-by: Greg Kroah-Hartman

usb: chipidea: udc: don't do hardware access if gadget has stopped

2019-08-21T16:45:49Z

After _gadget_stop_activity is executed, we can consider the hardware operation for gadget has finished, and the udc can be stopped and enter low power mode. So, any later hardware operations (from usb_ep_ops APIs or usb_gadget_ops APIs) should be considered invalid, any deinitializatons has been covered at _gadget_stop_activity. I meet this problem when I plug out usb cable from PC using mass_storage gadget, my callstack like: vbus interrupt->.vbus_session-> composite_disconnect ->pm_runtime_put_sync(&_gadget->dev), the composite_disconnect will call fsg_disable, but fsg_disable calls usb_ep_disable using async way, there are register accesses for usb_ep_disable. So sometimes, I get system hang due to visit register without clock, sometimes not. The Linux Kernel USB maintainer Alan Stern suggests this kinds of solution. See: http://marc.info/?l=linux-usb&m=138541769810983&w=2. Cc: #v4.9+ Signed-off-by: Peter Chen Link: https://lore.kernel.org/r/20190820020503.27080-2-peter.chen@nxp.com Signed-off-by: Greg Kroah-Hartman