linux-stable.git/drivers/thunderbolt/switch.c, branch v5.12.2

thunderbolt: Initialize HopID IDAs in tb_switch_alloc()

2021-03-08T11:22:35+00:00

If there is a failure before the tb_switch_add() is called the switch
object is released by tb_switch_release() but at that point HopID IDAs
have not yet been initialized. So we see splat like this:

BUG: spinlock bad magic on CPU#2, kworker/u8:5/115
...
Workqueue: thunderbolt0 tb_handle_hotplug
Call Trace:
 dump_stack+0x97/0xdc
 ? spin_bug+0x9a/0xa7
 do_raw_spin_lock+0x68/0x98
 _raw_spin_lock_irqsave+0x3f/0x5d
 ida_destroy+0x4f/0x127
 tb_switch_release+0x6d/0xfd
 device_release+0x2c/0x7d
 kobject_put+0x9b/0xbc
 tb_handle_hotplug+0x278/0x452
 process_one_work+0x1db/0x396
 worker_thread+0x216/0x375
 kthread+0x14d/0x155
 ? pr_cont_work+0x58/0x58
 ? kthread_blkcg+0x2e/0x2e
 ret_from_fork+0x1f/0x40

Fix this by always initializing HopID IDAs in tb_switch_alloc().

Fixes: 0b2863ac3cfd ("thunderbolt: Add functions for allocating and releasing HopIDs")
Cc: stable@vger.kernel.org
Reported-by: Chiranjeevi Rapolu 
Signed-off-by: Mika Westerberg

thunderbolt: Add support for PCIe tunneling disabled (SL5)

2021-02-04T07:45:24+00:00

Recent Intel Thunderbolt firmware connection manager has support for
another security level, SL5, that disables PCIe tunneling. This option
can be turned on from the BIOS.

When this is set the driver exposes a new security level "nopcie" to the
userspace and hides the authorized attribute under connected devices.

While there we also hide it when "dponly" security level is enabled
since it is not really usable in that case anyway.

Signed-off-by: Mika Westerberg 
Acked-by: Yehezkel Bernat

thunderbolt: switch: Fix kernel-doc descriptions of non-static functions

2021-02-04T07:38:38+00:00

Fix kernel-doc descriptions of all non-static functions. This also gets
rid of the warnings on W=1 build.

Reported-by: Lee Jones 
Signed-off-by: Mika Westerberg 
Reviewed-by: Lee Jones

thunderbolt: switch: Fix function name in the header

2021-01-28T10:57:06+00:00

Fixes the following W=1 kernel build warning(s):

 drivers/thunderbolt/switch.c:1322: warning: expecting prototype for reset_switch(). Prototype was for tb_switch_reset() instead

Cc: Andreas Noever 
Cc: Michael Jamet 
Cc: Yehezkel Bernat 
Cc: linux-usb@vger.kernel.org
Signed-off-by: Lee Jones 
Signed-off-by: Mika Westerberg

thunderbolt: switch: Demote a bunch of non-conformant kernel-doc headers

2021-01-28T10:52:24+00:00

Fixes the following W=1 kernel build warning(s):

 drivers/thunderbolt/switch.c:730: warning: Function parameter or member 'port' not described in 'tb_init_port'
 drivers/thunderbolt/switch.c:1348: warning: Function parameter or member 'sw' not described in 'tb_plug_events_active'
 drivers/thunderbolt/switch.c:1348: warning: Function parameter or member 'active' not described in 'tb_plug_events_active'

Cc: Andreas Noever 
Cc: Michael Jamet 
Cc: Yehezkel Bernat 
Cc: linux-usb@vger.kernel.org
Signed-off-by: Lee Jones 
[ mw: Demote only static functions ]
Signed-off-by: Mika Westerberg

thunderbolt: Add support for de-authorizing devices

2021-01-15T07:53:06+00:00

In some cases it is useful to be able de-authorize devices. For example
if user logs out the userspace can have a policy that disconnects PCIe
devices until logged in again. This is only possible for software based
connection manager as it directly controls the tunnels.

For this reason make the authorized attribute accept writing 0 which
makes the software connection manager to tear down the corresponding
PCIe tunnel. Userspace can check if this is supported by reading a new
domain attribute deauthorization, that holds 1 in that case.

While there correct tb_domain_approve_switch() kernel-doc and
description of authorized attribute to mention that it is only about
PCIe tunnels.

Cc: Christian Kellner 
Signed-off-by: Mika Westerberg 
Acked-by: Yehezkel Bernat

thunderbolt: Start lane initialization after sleep

2021-01-11T14:15:42+00:00

USB4 spec says that for TBT3 compatible device routers the connection
manager needs to set SLI (Start Lane Initialization) to get the lanes
that were not connected back to functional state after sleep. Same needs
to be done if the link was XDomain.

Signed-off-by: Mika Westerberg 
Acked-by: Yehezkel Bernat

thunderbolt: Constify static attribute_group structs

2021-01-11T14:13:48+00:00

The only usage of these is to put their addresses in arrays of pointers
to const attribute_groups. Make them const to allow the compiler to put
them in read-only memory.

Signed-off-by: Rikard Falkeborn 
Signed-off-by: Mika Westerberg

thunderbolt: Perform USB4 router NVM upgrade in two phases

2020-11-30T11:39:24+00:00

The currect code expects that the router returns back the status of the
NVM authentication immediately. When tested against a real USB4 device
what happens is that the router is reset and only after that the result
is updated in the ROUTER_CS_26 register status field. This also seems to
align better what the spec suggests.

For this reason do the same what we already do with the Thunderbolt 3
devices and perform the NVM upgrade in two phases. First start the
NVM_AUTH router operation and once the router is added back after the
reset read the status in ROUTER_CS_26 and expose it to the userspace
accordingly.

Signed-off-by: Mika Westerberg

thunderbolt: Add functions for enabling and disabling lane bonding on XDomain

2020-11-11T07:20:16+00:00

These can be used by service drivers to enable and disable lane bonding
as needed.

Signed-off-by: Isaac Hazan 
Signed-off-by: Mika Westerberg 
Acked-by: Yehezkel Bernat 
Reviewed-by: Greg Kroah-Hartman