linux-next.git/drivers/usb, branch master

Merge branch 'usb-next' of https://git.kernel.org/pub/scm/linux/kernel/git/johan/usb-serial.git

2026-07-17T17:16:49+00:00

Merge branch 'usb-next' of https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb.git

2026-07-17T17:16:45+00:00

Merge branch 'driver-core-next' of https://git.kernel.org/pub/scm/linux/kernel/git/driver-core/driver-core.git

2026-07-17T17:16:44+00:00

# Conflicts: # include/linux/platform_device.h

Merge branch 'usb-linus' of https://git.kernel.org/pub/scm/linux/kernel/git/johan/usb-serial.git

2026-07-17T12:20:07+00:00

usb: atm: ueagle-atm: reject descriptors that confuse probe and disconnect

2026-07-17T10:28:37+00:00

uea_probe() distinguishes a pre-firmware device from a post-firmware one using the USB id (UEA_IS_PREFIRM()), and stores a different object as the interface data in each case: a 'struct completion' for a pre-firmware device (to be waited on in .disconnect()), or a 'struct usbatm_data' for a post-firmware one. uea_disconnect() instead tells the two apart by the number of interfaces of the active configuration (a pre-firmware device exposes a single interface, ADI930 has 2 and eagle has 3), and casts the interface data accordingly. Because the two handlers use different criteria, a crafted device that advertises a pre-firmware id together with a multi-interface descriptor (or a post-firmware id with a single interface) makes them disagree: the small 'struct completion' stored by uea_probe() is then passed to usbatm_usb_disconnect(), which casts it to 'struct usbatm_data' and takes instance->serialize, reading past the end of the allocation: BUG: KASAN: slab-out-of-bounds in __mutex_lock+0x152a/0x1b80 Read of size 8 at addr ffff8880470e2c60 by task kworker/1:2/982 ... __mutex_lock+0x152a/0x1b80 usbatm_usb_disconnect+0x70/0x820 uea_disconnect+0x133/0x2c0 usb_unbind_interface+0x1dd/0x9e0 ... which belongs to the cache kmalloc-96 of size 96 The buggy address is located 0 bytes to the right of allocated 96-byte region [ffff8880470e2c00, ffff8880470e2c60) Reject such inconsistent descriptors in uea_probe() so that both handlers always make the same pre/post-firmware decision. Reported-by: syzbot+e62a973f8322b3bbe3ac@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=e62a973f8322b3bbe3ac Fixes: e2674dfbed8a ("usb: atm: ueagle-atm: wait for pre-firmware load in .disconnect()") Signed-off-by: Diego Fernando Mancera Gomez Acked-by: Stanislaw Gruszka Link: https://patch.msgid.link/20260717080704.1264-1-diegomancera.dev@gmail.com Signed-off-by: Greg Kroah-Hartman

usb: typec: ucsi: yoga_c630: Remove redundant duplicate altmode handling

2026-07-16T15:14:34+00:00

This reverts commit e0c48e42d818 ("usb: typec: ucsi: yoga-c630: remove duplicate AltModes"). The yoga_c630 driver previously implemented its own duplicate altmode detection in yoga_c630_ucsi_update_altmodes() to work around buggy EC firmware that returns duplicate AltModes instead of empty ones. With the introduction of the common ucsi_altmode_is_duplicate() helper in both the standard and nvidia registration paths, duplicate detection is now handled automatically in the core UCSI code. This makes the yoga_c630-specific implementation added in commit e0c48e42d818 ("usb: typec: ucsi: yoga-c630: remove duplicate AltModes") redundant. Remove yoga_c630_ucsi_update_altmodes() and its callback to eliminate code duplication and simplify the driver. Note that this causes the driver to switch back from the nvidia registration path to the standard path, which is the original behavior before commit e0c48e42d818 ("usb: typec: ucsi: yoga-c630: remove duplicate AltModes"). Both paths now include duplicate detection, ensuring the firmware bug is still properly handled. Signed-off-by: Chia-Lin Kao (AceLan) Link: https://patch.msgid.link/20260713084323.287516-3-acelan.kao@canonical.com Signed-off-by: Greg Kroah-Hartman

usb: typec: ucsi: Add duplicate detection to nvidia registration path

2026-07-16T15:14:27+00:00

Extend the duplicate altmode detection to ucsi_register_altmodes_nvidia() which is used when a driver provides the update_altmodes() callback. This ensures all drivers benefit from duplicate detection, whether they use the standard registration path or the nvidia path with update_altmodes callback. Without this fix, drivers using the nvidia path (like yoga_c630) would still encounter duplicate altmode registration errors from buggy firmware. Fixes: a79f16efcd00 ("usb: typec: ucsi: Add support for the partner USB Modes") Cc: stable Signed-off-by: Chia-Lin Kao (AceLan) Link: https://patch.msgid.link/20260713084323.287516-2-acelan.kao@canonical.com Signed-off-by: Greg Kroah-Hartman

usb: typec: ucsi: Detect and skip duplicate altmodes from buggy firmware

2026-07-16T15:14:18+00:00

Some firmware implementations incorrectly return the same altmode multiple times at different offsets when queried via UCSI_GET_ALTERNATE_MODES. This causes sysfs duplicate filename errors and kernel call traces when the driver attempts to register the same altmode twice: sysfs: cannot create duplicate filename '/devices/.../typec/port0/port0.0/partner' typec-thunderbolt port0-partner.1: failed to create symlinks typec-thunderbolt port0-partner.1: probe with driver typec-thunderbolt failed with error -17 The matching rules differ by recipient: - UCSI_RECIPIENT_CON (port) and UCSI_RECIPIENT_SOP_P (plug): Two altmodes with identical SVID and VDO are byte-for-byte duplicates and the second has no observable function, so drop it. - UCSI_RECIPIENT_SOP (partner): The typec class binds each partner altmode to a port altmode of the same SVID via altmode_match()/device_find_child(), which returns the first port altmode with a matching SVID. If the partner advertises more altmodes for SVID X than the port advertises, the surplus partner altmode(s) collapse onto an already-paired port altmode and trigger the "duplicate filename .../partner" sysfs error during typec_altmode_create_links(). Use the port-side altmode count for SVID X as the authoritative cap and reject any partner altmode that would exceed it. This preserves legitimate multi-Mode partner altmodes (vendor SVIDs that the port really does advertise more than once) while filtering the firmware-generated duplicates that have no port counterpart, and is therefore stricter than a plain SVID+VDO comparison (which still admits the Thunderbolt case where firmware reports the same SVID twice with different VDOs) without being over-broad like a plain SVID match (which would falsely drop legitimate vendor multi-Mode entries). If a duplicate is detected, skip it and emit a clean warning instead of generating a kernel call trace: ucsi_acpi USBC000:00: con2: Firmware bug: duplicate partner altmode SVID 0x8087 at offset 1, ignoring. ucsi_acpi USBC000:00: con2: VDO mismatch: 0x8087a043 vs 0x00000001 The duplicate detection logic lives in a reusable helper ucsi_altmode_is_duplicate() and is invoked from ucsi_register_altmodes(). It applies to all three recipient types: partner (SOP), port (CON), and plug (SOP_P) altmodes. Fixes: a79f16efcd00 ("usb: typec: ucsi: Add support for the partner USB Modes") Cc: stable Signed-off-by: Chia-Lin Kao (AceLan) unchanged: still SVID+VDO exact-dup match. Reviewed-by: Heikki Krogerus Link: https://patch.msgid.link/20260713084323.287516-1-acelan.kao@canonical.com Signed-off-by: Greg Kroah-Hartman

usb: gadget: dummy_hcd: prevent fifo_req reuse during giveback

2026-07-16T15:12:36+00:00

dummy_hcd embeds a single shared usb_request (dum->fifo_req) that the "emulated single-request FIFO" fast-path in dummy_queue() reuses for small IN transfers: it copies the caller's request into it (req->req = *_req) and queues it, treating list_empty(&fifo_req.queue) as "the slot is free". The completion side (dummy_timer/transfer/nuke/dummy_dequeue) follows the standard pattern: list_del_init(&req->queue) unlinks the request, then the lock is dropped and usb_gadget_giveback_request() invokes req->complete(). But list_del_init() makes fifo_req.queue look empty *before* the completion callback returns, so a concurrent dummy_queue() on another CPU sees the slot as free, reuses fifo_req and runs req->req = *_req -- overwriting req->complete while dummy_timer is mid-calling it. The indirect call then jumps to a clobbered pointer, causing a general protection fault / page fault in dummy_timer (syzkaller extid faf3a6cf579fc65591ca). The clobbering write is an in-bounds memcpy on a live shared object, so KASAN cannot flag it. Add a fifo_req_busy bit covering the shared request's whole lifetime: set it in dummy_queue() when the FIFO fast-path takes fifo_req (making it the fast-path guard, replacing the list_empty(&fifo_req.queue) test), and clear it after the completion callback has returned, via a dummy_giveback() helper used at all four gadget-request giveback sites. The shared slot can no longer be reused until its completion callback has finished. Reported-by: syzbot+faf3a6cf579fc65591ca@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=faf3a6cf579fc65591ca Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable Signed-off-by: Jinchao Wang Reviewed-by: Alan Stern Link: https://patch.msgid.link/5db8bba5b3499a86cd2e776f9918126b68b2508b.1784198306.git.wangjinchao600@gmail.com Signed-off-by: Greg Kroah-Hartman

usb: chipidea: fix usage_count leak when autosuspend_delay is negative

2026-07-16T15:12:21+00:00

The probe() calls pm_runtime_use_autosuspend(), but remove() does not call pm_runtime_dont_use_autosuspend(). This can lead to a usage_count leak if autosuspend_delay is set to a negative value. The pm_runtime_use_autosuspend() also notes that it's important to undo this with pm_runtime_dont_use_autosuspend() at driver exit time. Fixes: 1f874edcb731 ("usb: chipidea: add runtime power management support") Cc: stable Assisted-by: Claude:claude-sonnet-4.6 Signed-off-by: Xu Yang Reviewed-by: Frank Li Link: https://patch.msgid.link/20260716104126.2763454-1-xu.yang_2@oss.nxp.com Signed-off-by: Greg Kroah-Hartman