kernel.git/drivers/usb/gadget/function/f_fs.c, branch linux-4.9.y

usb: gadget: f_fs: Use stream_open() for endpoint files

2022-01-27T07:47:36Z

[ Upstream commit c76ef96fc00eb398c8fc836b0eb2f82bcc619dc7 ] Function fs endpoint file operations are synchronized via an interruptible mutex wait. However we see threads that do ep file operations concurrently are getting blocked for the mutex lock in __fdget_pos(). This is an uninterruptible wait and we see hung task warnings and kernel panic if hung_task_panic systcl is enabled if host does not send/receive the data for long time. The reason for threads getting blocked in __fdget_pos() is due to the file position protection introduced by the commit 9c225f2655e3 ("vfs: atomic f_pos accesses as per POSIX"). Since function fs endpoint files does not have the notion of the file position, switch to the stream mode. This will bypass the file position mutex and threads will be blocked in interruptible state for the function fs mutex. It should not affects user space as we are only changing the task state changes the task state from UNINTERRUPTIBLE to INTERRUPTIBLE while waiting for the USB transfers to be finished. However there is a slight change to the O_NONBLOCK behavior. Earlier threads that are using O_NONBLOCK are also getting blocked inside fdget_pos(). Now they reach to function fs and error code is returned. The non blocking behavior is actually honoured now. Reviewed-by: John Keeping Signed-off-by: Pavankumar Kondeti Link: https://lore.kernel.org/r/1636712682-1226-1-git-send-email-quic_pkondeti@quicinc.com Signed-off-by: Greg Kroah-Hartman Signed-off-by: Sasha Levin

usb: gadget: f_fs: Clear ffs_eventfd in ffs_data_clear.

2022-01-05T11:31:25Z

commit b1e0887379422975f237d43d8839b751a6bcf154 upstream. ffs_data_clear is indirectly called from both ffs_fs_kill_sb and ffs_ep0_release, so it ends up being called twice when userland closes ep0 and then unmounts f_fs. If userland provided an eventfd along with function's USB descriptors, it ends up calling eventfd_ctx_put as many times, causing a refcount underflow. NULL-ify ffs_eventfd to prevent these extraneous eventfd_ctx_put calls. Also, set epfiles to NULL right after de-allocating it, for readability. For completeness, ffs_data_clear actually ends up being called thrice, the last call being before the whole ffs structure gets freed, so when this specific sequence happens there is a second underflow happening (but not being reported): /sys/kernel/debug/tracing# modprobe usb_f_fs /sys/kernel/debug/tracing# echo ffs_data_clear > set_ftrace_filter /sys/kernel/debug/tracing# echo function > current_tracer /sys/kernel/debug/tracing# echo 1 > tracing_on (setup gadget, run and kill function userland process, teardown gadget) /sys/kernel/debug/tracing# echo 0 > tracing_on /sys/kernel/debug/tracing# cat trace smartcard-openp-436 [000] ..... 1946.208786: ffs_data_clear <-ffs_data_closed smartcard-openp-431 [000] ..... 1946.279147: ffs_data_clear <-ffs_data_closed smartcard-openp-431 [000] .n... 1946.905512: ffs_data_clear <-ffs_data_put Warning output corresponding to above trace: [ 1946.284139] WARNING: CPU: 0 PID: 431 at lib/refcount.c:28 refcount_warn_saturate+0x110/0x15c [ 1946.293094] refcount_t: underflow; use-after-free. [ 1946.298164] Modules linked in: usb_f_ncm(E) u_ether(E) usb_f_fs(E) hci_uart(E) btqca(E) btrtl(E) btbcm(E) btintel(E) bluetooth(E) nls_ascii(E) nls_cp437(E) vfat(E) fat(E) bcm2835_v4l2(CE) bcm2835_mmal_vchiq(CE) videobuf2_vmalloc(E) videobuf2_memops(E) sha512_generic(E) videobuf2_v4l2(E) sha512_arm(E) videobuf2_common(E) videodev(E) cpufreq_dt(E) snd_bcm2835(CE) brcmfmac(E) mc(E) vc4(E) ctr(E) brcmutil(E) snd_soc_core(E) snd_pcm_dmaengine(E) drbg(E) snd_pcm(E) snd_timer(E) snd(E) soundcore(E) drm_kms_helper(E) cec(E) ansi_cprng(E) rc_core(E) syscopyarea(E) raspberrypi_cpufreq(E) sysfillrect(E) sysimgblt(E) cfg80211(E) max17040_battery(OE) raspberrypi_hwmon(E) fb_sys_fops(E) regmap_i2c(E) ecdh_generic(E) rfkill(E) ecc(E) bcm2835_rng(E) rng_core(E) vchiq(CE) leds_gpio(E) libcomposite(E) fuse(E) configfs(E) ip_tables(E) x_tables(E) autofs4(E) ext4(E) crc16(E) mbcache(E) jbd2(E) crc32c_generic(E) sdhci_iproc(E) sdhci_pltfm(E) sdhci(E) [ 1946.399633] CPU: 0 PID: 431 Comm: smartcard-openp Tainted: G C OE 5.15.0-1-rpi #1 Debian 5.15.3-1 [ 1946.417950] Hardware name: BCM2835 [ 1946.425442] Backtrace: [ 1946.432048] [] (dump_backtrace) from [] (show_stack+0x20/0x24) [ 1946.448226] r7:00000009 r6:0000001c r5:c04a948c r4:c0a64e2c [ 1946.458412] [] (show_stack) from [] (dump_stack+0x28/0x30) [ 1946.470380] [] (dump_stack) from [] (__warn+0xe8/0x154) [ 1946.482067] r5:c04a948c r4:c0a71dc8 [ 1946.490184] [] (__warn) from [] (warn_slowpath_fmt+0xa0/0xe4) [ 1946.506758] r7:00000009 r6:0000001c r5:c0a71dc8 r4:c0a71e04 [ 1946.517070] [] (warn_slowpath_fmt) from [] (refcount_warn_saturate+0x110/0x15c) [ 1946.535309] r8:c0100224 r7:c0dfcb84 r6:ffffffff r5:c3b84c00 r4:c24a17c0 [ 1946.546708] [] (refcount_warn_saturate) from [] (eventfd_ctx_put+0x48/0x74) [ 1946.564476] [] (eventfd_ctx_put) from [] (ffs_data_clear+0xd0/0x118 [usb_f_fs]) [ 1946.582664] r5:c3b84c00 r4:c2695b00 [ 1946.590668] [] (ffs_data_clear [usb_f_fs]) from [] (ffs_data_closed+0x9c/0x150 [usb_f_fs]) [ 1946.609608] r5:bf54d014 r4:c2695b00 [ 1946.617522] [] (ffs_data_closed [usb_f_fs]) from [] (ffs_fs_kill_sb+0x2c/0x30 [usb_f_fs]) [ 1946.636217] r7:c0dfcb84 r6:c3a12260 r5:bf54d014 r4:c229f000 [ 1946.646273] [] (ffs_fs_kill_sb [usb_f_fs]) from [] (deactivate_locked_super+0x54/0x9c) [ 1946.664893] r5:bf54d014 r4:c229f000 [ 1946.672921] [] (deactivate_locked_super) from [] (deactivate_super+0x60/0x64) [ 1946.690722] r5:c2a09000 r4:c229f000 [ 1946.698706] [] (deactivate_super) from [] (cleanup_mnt+0xe4/0x14c) [ 1946.715553] r5:c2a09000 r4:00000000 [ 1946.723528] [] (cleanup_mnt) from [] (__cleanup_mnt+0x1c/0x20) [ 1946.739922] r7:c0dfcb84 r6:c3a12260 r5:c3a126fc r4:00000000 [ 1946.750088] [] (__cleanup_mnt) from [] (task_work_run+0x84/0xb8) [ 1946.766602] [] (task_work_run) from [] (do_work_pending+0x470/0x56c) [ 1946.783540] r7:5ac3c35a r6:c0d0424c r5:c200bfb0 r4:c200a000 [ 1946.793614] [] (do_work_pending) from [] (slow_work_pending+0xc/0x20) [ 1946.810553] Exception stack(0xc200bfb0 to 0xc200bff8) [ 1946.820129] bfa0: 00000000 00000000 000000aa b5e21430 [ 1946.837104] bfc0: bef867a0 00000001 bef86840 00000034 bef86838 bef86790 bef86794 bef867a0 [ 1946.854125] bfe0: 00000000 bef86798 b67b7a1c b6d626a4 60000010 b5a23760 [ 1946.865335] r10:00000000 r9:c200a000 r8:c0100224 r7:00000034 r6:bef86840 r5:00000001 [ 1946.881914] r4:bef867a0 [ 1946.888793] ---[ end trace 7387f2a9725b28d0 ]--- Fixes: 5e33f6fdf735 ("usb: gadget: ffs: add eventfd notification about ffs events") Cc: stable Signed-off-by: Vincent Pelletier Link: https://lore.kernel.org/r/f79eeea29f3f98de6782a064ec0f7351ad2f598f.1639793920.git.plr.vincent@gmail.com Signed-off-by: Greg Kroah-Hartman

usb: gadget/function/f_fs string table fix for multiple languages

2021-05-22T08:40:20Z

commit 55b74ce7d2ce0b0058f3e08cab185a0afacfe39e upstream. Fixes bug with the handling of more than one language in the string table in f_fs.c. str_count was not reset for subsequent language codes. str_count-- "rolls under" and processes u32 max strings on the processing of the second language entry. The existing bug can be reproduced by adding a second language table to the structure "strings" in tools/usb/ffs-test.c. Signed-off-by: Dean Anderson Link: https://lore.kernel.org/r/20210317224109.21534-1-dean@sensoray.com Cc: stable Signed-off-by: Greg Kroah-Hartman

usb: gadget: f_fs: Re-use SS descriptors for SuperSpeedPlus

2020-12-29T12:44:51Z

commit a353397b0d5dfa3c99b372505db3378fc919c6c6 upstream. In many cases a function that supports SuperSpeed can very well operate in SuperSpeedPlus, if a gadget controller supports it, as the endpoint descriptors (and companion descriptors) are generally identical and can be re-used. This is true for two commonly used functions: Android's ADB and MTP. So we can simply assign the usb_function's ssp_descriptors array to point to its ss_descriptors, if available. Similarly, we need to allow an epfile's ioctl for FUNCTIONFS_ENDPOINT_DESC to correctly return the corresponding SuperSpeed endpoint descriptor in case the connected speed is SuperSpeedPlus as well. The only exception is if a function wants to implement an Isochronous endpoint capable of transferring more than 48KB per service interval when operating at greater than USB 3.1 Gen1 speed, in which case it would require an additional SuperSpeedPlus Isochronous Endpoint Companion descriptor to be returned as part of the Configuration Descriptor. Support for that would need to be separately added to the userspace-facing FunctionFS API which may not be a trivial task--likely a new descriptor format (v3?) may need to be devised to allow for separate SS and SSP descriptors to be supplied. Signed-off-by: Jack Pham Cc: stable Link: https://lore.kernel.org/r/20201027230731.9073-1-jackp@codeaurora.org Signed-off-by: Greg Kroah-Hartman Signed-off-by: Greg Kroah-Hartman

usb: gadget: f_fs: Use local copy of descriptors for userspace copy

2020-12-11T12:37:59Z

commit a4b98a7512f18534ce33a7e98e49115af59ffa00 upstream. The function may be unbound causing the ffs_ep and its descriptors to be freed while userspace is in the middle of an ioctl requesting the same descriptors. Avoid dangling pointer reference by first making a local copy of desctiptors before releasing the spinlock. Fixes: c559a3534109 ("usb: gadget: f_fs: add ioctl returning ep descriptor") Reviewed-by: Peter Chen Signed-off-by: Vamsi Krishna Samavedam Signed-off-by: Jack Pham Cc: stable Link: https://lore.kernel.org/r/20201130203453.28154-1-jackp@codeaurora.org Signed-off-by: Greg Kroah-Hartman

usb: f_fs: Clear OS Extended descriptor counts to zero in ffs_data_reset()

2020-05-02T15:23:14Z

commit 1c2e54fbf1da5e5445a0ab132c862b02ccd8d230 upstream. For userspace functions using OS Descriptors, if a function also supplies Extended Property descriptors currently the counts and lengths stored in the ms_os_descs_ext_prop_{count,name_len,data_len} variables are not getting reset to 0 during an unbind or when the epfiles are closed. If the same function is re-bound and the descriptors are re-written, this results in those count/length variables to monotonically increase causing the VLA allocation in _ffs_func_bind() to grow larger and larger at each bind/unbind cycle and eventually fail to allocate. Fix this by clearing the ms_os_descs_ext_prop count & lengths to 0 in ffs_data_reset(). Fixes: f0175ab51993 ("usb: gadget: f_fs: OS descriptors support") Cc: stable@vger.kernel.org Signed-off-by: Udipto Goswami Signed-off-by: Sriharsha Allenki Reviewed-by: Manu Gautam Link: https://lore.kernel.org/r/20200402044521.9312-1-sallenki@codeaurora.org Signed-off-by: Greg Kroah-Hartman

usb: gadget: f_fs: Fix use after free issue as part of queue failure

2020-04-24T05:58:51Z

commit f63ec55ff904b2f2e126884fcad93175f16ab4bb upstream. In AIO case, the request is freed up if ep_queue fails. However, io_data->req still has the reference to this freed request. In the case of this failure if there is aio_cancel call on this io_data it will lead to an invalid dequeue operation and a potential use after free issue. Fix this by setting the io_data->req to NULL when the request is freed as part of queue failure. Fixes: 2e4c7553cd6f ("usb: gadget: f_fs: add aio support") Signed-off-by: Sriharsha Allenki CC: stable Reviewed-by: Peter Chen Link: https://lore.kernel.org/r/20200326115620.12571-1-sallenki@codeaurora.org Signed-off-by: Greg Kroah-Hartman

usb: gadget: ffs: ffs_aio_cancel(): Save/restore IRQ flags

2020-03-11T06:53:09Z

[ Upstream commit 43d565727a3a6fd24e37c7c2116475106af71806 ] ffs_aio_cancel() can be called from both interrupt and thread context. Make sure that the current IRQ state is saved and restored by using spin_{un,}lock_irq{save,restore}(). Otherwise undefined behavior might occur. Acked-by: Michal Nazarewicz Signed-off-by: Lars-Peter Clausen Signed-off-by: Alexandru Ardelean Signed-off-by: Felipe Balbi Signed-off-by: Sasha Levin

usb: gadget: Zero ffs_io_data

2019-08-04T07:33:38Z

[ Upstream commit 508595515f4bcfe36246e4a565cf280937aeaade ] In some cases the "Allocate & copy" block in ffs_epfile_io() is not executed. Consequently, in such a case ffs_alloc_buffer() is never called and struct ffs_io_data is not initialized properly. This in turn leads to problems when ffs_free_buffer() is called at the end of ffs_epfile_io(). This patch uses kzalloc() instead of kmalloc() in the aio case and memset() in non-aio case to properly initialize struct ffs_io_data. Signed-off-by: Andrzej Pietrasiewicz Signed-off-by: Felipe Balbi Signed-off-by: Sasha Levin

usb: f_fs: Avoid crash due to out-of-scope stack ptr access

2019-04-05T20:29:11Z

[ Upstream commit 54f64d5c983f939901dacc8cfc0983727c5c742e ] Since the 5.0 merge window opened, I've been seeing frequent crashes on suspend and reboot with the trace: [ 36.911170] Unable to handle kernel paging request at virtual address ffffff801153d660 [ 36.912769] Unable to handle kernel paging request at virtual address ffffff800004b564 ... [ 36.950666] Call trace: [ 36.950670] queued_spin_lock_slowpath+0x1cc/0x2c8 [ 36.950681] _raw_spin_lock_irqsave+0x64/0x78 [ 36.950692] complete+0x28/0x70 [ 36.950703] ffs_epfile_io_complete+0x3c/0x50 [ 36.950713] usb_gadget_giveback_request+0x34/0x108 [ 36.950721] dwc3_gadget_giveback+0x50/0x68 [ 36.950723] dwc3_thread_interrupt+0x358/0x1488 [ 36.950731] irq_thread_fn+0x30/0x88 [ 36.950734] irq_thread+0x114/0x1b0 [ 36.950739] kthread+0x104/0x130 [ 36.950747] ret_from_fork+0x10/0x1c I isolated this down to in ffs_epfile_io(): https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/usb/gadget/function/f_fs.c#n1065 Where the completion done is setup on the stack: DECLARE_COMPLETION_ONSTACK(done); Then later we setup a request and queue it, and wait for it: if (unlikely(wait_for_completion_interruptible(&done))) { /* * To avoid race condition with ffs_epfile_io_complete, * dequeue the request first then check * status. usb_ep_dequeue API should guarantee no race * condition with req->complete callback. */ usb_ep_dequeue(ep->ep, req); interrupted = ep->status < 0; } The problem is, that we end up being interrupted, dequeue the request, and exit. But then the irq triggers and we try calling complete() on the context pointer which points to now random stack space, which results in the panic. Alan Stern pointed out there is a bug here, in that the snippet above "assumes that usb_ep_dequeue() waits until the request has been completed." And that: wait_for_completion(&done); Is needed right after the usb_ep_dequeue(). Thus this patch implements that change. With it I no longer see the crashes on suspend or reboot. This issue seems to have been uncovered by behavioral changes in the dwc3 driver in commit fec9095bdef4e ("usb: dwc3: gadget: remove wait_end_transfer"). Cc: Alan Stern Cc: Felipe Balbi Cc: Zeng Tao Cc: Jack Pham Cc: Thinh Nguyen Cc: Chen Yu Cc: Jerry Zhang Cc: Lars-Peter Clausen Cc: Vincent Pelletier Cc: Andrzej Pietrasiewicz Cc: Greg Kroah-Hartman Cc: Linux USB List Suggested-by: Alan Stern Signed-off-by: John Stultz Signed-off-by: Felipe Balbi Signed-off-by: Sasha Levin