Merge branch 'next' into for-linus

Prepare input updates for 6.3 merge window.

Merge drm/drm-next into drm-misc-next

Backmerging into drm-misc-next to get DRM accelerator infrastructure,
which is required by ipuv driver.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>

Merge wireless into wireless-next

Due to the two cherry picked commits from wireless to wireless-next we have
several conflicts in mt76. To avoid any bugs with conflicts merge wireless into
wireless

Merge wireless into wireless-next

Due to the two cherry picked commits from wireless to wireless-next we have
several conflicts in mt76. To avoid any bugs with conflicts merge wireless into
wireless-next.

96f134dc1964 wifi: mt76: handle possible mt76_rx_token_consume failures
fe13dad8992b wifi: mt76: dma: do not increment queue head if mt76_dma_add_buf fails

show more ...

Merge tag 'v6.2-rc3' into next

Merge with mainline to bring in timer_shutdown_sync() API.

Merge drm/drm-fixes into drm-misc-fixes

Let's start the fixes cycle.

Signed-off-by: Maxime Ripard <maxime@cerno.tech>

Merge drm/drm-next into drm-intel-next

Sync up with v6.2-rc1.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Merge drm/drm-next into drm-intel-gt-next

Sync after v6.2-rc1 landed in drm-next.

We need to get some dependencies in place before we can merge
the fixes series from Gwan-gyeong and Chris.

Referen

Merge drm/drm-next into drm-intel-gt-next

Sync after v6.2-rc1 landed in drm-next.

We need to get some dependencies in place before we can merge
the fixes series from Gwan-gyeong and Chris.

References: https://lore.kernel.org/all/Y6x5JCDnh2rvh4lA@intel.com/
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

show more ...

Merge tag 'v6.2-rc1' into media_tree

Linux 6.2-rc1

* tag 'v6.2-rc1': (14398 commits)
Linux 6.2-rc1
treewide: Convert del_timer*() to timer_shutdown*()
pstore: Properly assign mem_type propert

Merge tag 'v6.2-rc1' into media_tree

Linux 6.2-rc1

* tag 'v6.2-rc1': (14398 commits)
Linux 6.2-rc1
treewide: Convert del_timer*() to timer_shutdown*()
pstore: Properly assign mem_type property
pstore: Make sure CONFIG_PSTORE_PMSG selects CONFIG_RT_MUTEXES
cfi: Fix CFI failure with KASAN
perf python: Fix splitting CC into compiler and options
afs: Stop implementing ->writepage()
afs: remove afs_cache_netfs and afs_zap_permits() declarations
afs: remove variable nr_servers
afs: Fix lost servers_outstanding count
ALSA: usb-audio: Add new quirk FIXED_RATE for JBL Quantum810 Wireless
ALSA: azt3328: Remove the unused function snd_azf3328_codec_outl()
gcov: add support for checksum field
test_maple_tree: add test for mas_spanning_rebalance() on insufficient data
maple_tree: fix mas_spanning_rebalance() on insufficient data
hugetlb: really allocate vma lock for all sharable vmas
kmsan: export kmsan_handle_urb
kmsan: include linux/vmalloc.h
mm/mempolicy: fix memory leak in set_mempolicy_home_node system call
mm, mremap: fix mremap() expanding vma with addr inside vma
...

show more ...

Merge remote-tracking branch 'torvalds/master' into perf/core

To resolve a trivial merge conflict with c302378bc157f6a7 ("libbpf:
Hashmap interface update to allow both long and void* keys/values"),

Merge remote-tracking branch 'torvalds/master' into perf/core

To resolve a trivial merge conflict with c302378bc157f6a7 ("libbpf:
Hashmap interface update to allow both long and void* keys/values"),
where a function present upstream was removed in the perf tools
development tree.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

show more ...

Merge tag 'for-linus-2022121301' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid

Pull HID updates from Jiri Kosina:

- iio support for the MCP2221 HID driver (Matt Ranostay)

- support fo

Merge tag 'for-linus-2022121301' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid

Pull HID updates from Jiri Kosina:

- iio support for the MCP2221 HID driver (Matt Ranostay)

- support for more than one hinge sensor in hid-sensor-custom (Yauhen
Kharuzhy)

- PS DualShock 4 controller support (Roderick Colenbrander)

- XP-PEN Deco LW support (José Expósito)

- other assorted code cleanups and device ID/quirk addtions

* tag 'for-linus-2022121301' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid: (51 commits)
HID: logitech HID++: Send SwID in GetProtocolVersion
HID: hid-elan: use default remove for hid device
HID: hid-alps: use default remove for hid device
HID: hid-sensor-custom: set fixed size for custom attributes
HID: i2c: let RMI devices decide what constitutes wakeup event
HID: playstation: fix DualShock4 bluetooth CRC endian issue.
HID: playstation: fix DualShock4 bluetooth memory corruption bug.
HID: apple: Swap Control and Command keys on Apple keyboards
HID: intel-ish-hid: ishtp: remove variable rb_count
HID: uclogic: Standardize test name prefix
HID: hid-sensor-custom: Allow more than one hinge angle sensor
HID: ft260: fix 'cast to restricted' kernel CI bot warnings
HID: ft260: missed NACK from busy device
HID: ft260: fix a NULL pointer dereference in ft260_i2c_write
HID: ft260: wake up device from power saving mode
HID: ft260: missed NACK from big i2c read
HID: ft260: remove SMBus Quick command support
HID: ft260: skip unexpected HID input reports
HID: ft260: do not populate /dev/hidraw device
HID: ft260: improve i2c large reads performance
...

show more ...

Merge branch 'for-6.2/ft260' into for-linus

- fixes and performance improvements to the hid-ft260 driver (Michael Zaidman)

HID: ft260: fix 'cast to restricted' kernel CI bot warnings

Fix 'cast to restricted' sparse warnings reported by kernel test robot
in https://lore.kernel.org/all/202211021607.ssjymlKi-lkp@intel.com/

HID: ft260: fix 'cast to restricted' kernel CI bot warnings

Fix 'cast to restricted' sparse warnings reported by kernel test robot
in https://lore.kernel.org/all/202211021607.ssjymlKi-lkp@intel.com/

Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: missed NACK from busy device

When writing into a slow device like an EEPROM chip, the
controller may exit the busy state before the device releases
the bus. In this case, the ft260_xfer_

HID: ft260: missed NACK from busy device

When writing into a slow device like an EEPROM chip, the
controller may exit the busy state before the device releases
the bus. In this case, the ft260_xfer_status returns success
before the data transfer completion.

The patch fixes it by returning from the ft260_xfer_status()
with the "-EAGAIN" on both controller and bus busy status when
appropriate.

It does not apply to the i2c combined transactions when after
the write IO, the controller keeps the bus busy until the read
IO and then between reading IOs to ensure an atomic operation.

Co-developed-by: Germain Hebert <germain.hebert@ca.abb.com>
Signed-off-by: Germain Hebert <germain.hebert@ca.abb.com>
Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: fix a NULL pointer dereference in ft260_i2c_write

The zero-length passed into the ft260_i2c_write() triggered the
NULL pointer dereference in the debug message on data[0] access.
Since t

HID: ft260: fix a NULL pointer dereference in ft260_i2c_write

The zero-length passed into the ft260_i2c_write() triggered the
NULL pointer dereference in the debug message on data[0] access.
Since the controller does not support a write of zero length,
let's not allow it.

Before:

$ sudo i2ctransfer -y 13 w0@0x51
Killed

After:

$ sudo i2ctransfer -y 13 w0@0x51
Error: Sending messages failed: Invalid argument

Reported-by: Enrik Berkhan <Enrik.Berkhan@inka.de>
Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: wake up device from power saving mode

The FT260 can enter a power saving mode after being idle for longer
than 5 seconds.

When being woken up from power saving mode by an I2C write requ

HID: ft260: wake up device from power saving mode

The FT260 can enter a power saving mode after being idle for longer
than 5 seconds.

When being woken up from power saving mode by an I2C write request,
a possible NACK is not correctly reported by the controller. As a
workaround, the driver will issue an I2C status report two times in
ft260_xfer_status() after the chip has been idle for more than 5s.

Co-developed-by: Enrik Berkhan <Enrik.Berkhan@inka.de>
Signed-off-by: Enrik Berkhan <Enrik.Berkhan@inka.de>
Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: missed NACK from big i2c read

The FT260 controller does not return NACK when performing a big
read (of multiple hid reports size) from a non-existing device
or from the device responding

HID: ft260: missed NACK from big i2c read

The FT260 controller does not return NACK when performing a big
read (of multiple hid reports size) from a non-existing device
or from the device responding with NACK when it is not ready
to serve the request. However, it responds correctly with NACK
to a read of up to a single hid report size.

To overcome this issue, we split the muli-report read request
into a read of a single HID report of 60 bytes size and a
multi-report read.

Big read of 256 bytes with first read of 60 bytes:

$ sudo ./i2cperf -d 2 -o 2 -s 256 -r 0-0xff 1 0x50 -S

[ +5.633280] ft260_i2c_write_read: off 0x0 rlen 255 wlen 2
[ +0.000006] ft260_i2c_write: rep 0xd0 addr 0x50 off 0 len 2 wlen 2 flag 0x2 d[0] 0x0
[ +0.013205] ft260_xfer_status: bus_status 0x20, clock 100
[ +0.000007] ft260_i2c_read: rep 0xc2 addr 0x50 len 255 rlen 60 flag 0x3
[ +0.010932] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.004733] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000006] ft260_i2c_read: rep 0xc2 addr 0x50 len 195 rlen 128 flag 0x0
[ +0.012572] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.005789] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.003189] ft260_raw_event: i2c resp: rep 0xd1 len 8
[ +0.004092] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000010] ft260_i2c_read: rep 0xc2 addr 0x50 len 67 rlen 67 flag 0x4
[ +0.011688] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.004700] ft260_raw_event: i2c resp: rep 0xd1 len 7
[ +0.004858] ft260_xfer_status: bus_status 0x20, clock 100

Read from non-existing device at address 8. The first 60 read responded
with NACK.

$ sudo ./i2cperf -d 2 -o 2 -s 256 -r 0-0xff 1 0x8 -S
[Oct19 15:37] ft260_i2c_write_read: off 0x0 rlen 255 wlen 2
[ +0.000007] ft260_i2c_write: rep 0xd0 addr 0x8 off 0 len 2 wlen 2 flag 0x2 d[0] 0x0
[ +0.022820] ft260_xfer_status: bus_status 0x20, clock 100
[ +0.000007] ft260_i2c_read: rep 0xc2 addr 0x8 len 255 rlen 60 flag 0x3
[ +0.010658] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.005965] ft260_xfer_status: bus_status 0x46, clock 100 <-- NACK
[ +0.000009] ft260 0003:0403:6030.0004: i2c bus error: 0x46
[ +0.007784] ft260_i2c_reset: done

Co-developed-by: Enrik Berkhan <Enrik.Berkhan@inka.de>
Signed-off-by: Enrik Berkhan <Enrik.Berkhan@inka.de>
Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: remove SMBus Quick command support

The i2cdetect uses the SMBus Quick command by default to scan devices
on the I2C bus. The FT260 implements an I2C bus controller. The SMBus
is derived

HID: ft260: remove SMBus Quick command support

The i2cdetect uses the SMBus Quick command by default to scan devices
on the I2C bus. The FT260 implements an I2C bus controller. The SMBus
is derived from I2C, but there are several differences between the
specifications of the two buses in the areas of timing, protocols,
operation modes, and electrical characteristics.

One of the differences is that the I2C devices allow the slave not
to ACK its slave address, but SMBus requires it to always ACK it as
a mechanism to detect a detachable device’s presence on the bus.
Since FT260 is the I2C bus controller, it does not acknowledge the
SMBus Quick write command, which sends a single bit to the device at
the place of the RD/WR bit.

The ft260 driver attempted to mimic the SMBus Quick Write functionality
by writing a single byte as the SMBus Byte Write command does.

Usually, one byte in the SMBus Quick Write will be fine. However, it may
cause problems with devices with a control register at offset 0, like
i2c muxes, for example, when scanned with the i2cdetect utility.

The i2cdetect with the "-r" option uses the SMBus Read Byte command,
which is a reasonable workaround. To prevent the I2C bus from locking
at write-only devices (most notably clock chips at address 0x69), use
the "-r" option in conjunction with scanning range parameters.

This patch removes the SMBus Quick command support.

$ sudo i2cdetect -y 13
Warning: Can't use SMBus Quick Write command, will skip some addresses
0 1 2 3 4 5 6 7 8 9 a b c d e f
00:
10:
20:
30: -- -- -- -- -- -- -- --
40:
50: 50 51 -- -- -- -- -- -- -- -- -- -- -- -- -- --
60:
70:

$ sudo i2cdetect -y -r 13
0 1 2 3 4 5 6 7 8 9 a b c d e f
00: -- -- -- -- -- -- -- --
10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
50: 50 51 -- -- -- -- -- -- -- -- -- -- -- -- -- --
60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
70: -- -- -- -- -- -- -- --

Reported-by: Vince Asbridge <VAsbridge@sanblaze.com>
Reported-by: Stephen Shirron <SShirron@sanblaze.com>
Reported-by: Enrik Berkhan <Enrik.Berkhan@inka.de>
Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: skip unexpected HID input reports

The FT260 is not supposed to generate unexpected HID reports. However,
in theory, the unsolicited HID Input reports can be issued by a specially
crafted

HID: ft260: skip unexpected HID input reports

The FT260 is not supposed to generate unexpected HID reports. However,
in theory, the unsolicited HID Input reports can be issued by a specially
crafted malicious USB device masquerading as FT260 when the attacker has
physical access to the USB port. In this case, the read_buf pointer points
to the final data portion of the previous I2C Read transfer, and the memcpy
invoked in the ft260_raw_event() will try copying the content of the
unexpected report into the wrong location.

This commit sets the Read buffer pointer to NULL on the I2C Read
transaction completion and checks it in the ft260_raw_event() to detect
and skip the unsolicited Input report.

Reported-by: Enrik Berkhan <Enrik.Berkhan@inka.de>
Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: do not populate /dev/hidraw device

Do not populate the /dev/hidraw on ft260 interfaces when the hid-ft260
driver is loaded.

$ sudo insmod hid-ft260.ko
$ ls /dev/hidraw*
/dev/hidraw0

$

HID: ft260: do not populate /dev/hidraw device

Do not populate the /dev/hidraw on ft260 interfaces when the hid-ft260
driver is loaded.

$ sudo insmod hid-ft260.ko
$ ls /dev/hidraw*
/dev/hidraw0

$ sudo rmmod hid-ft260.ko
$ ls /dev/hidraw*
/dev/hidraw0 /dev/hidraw1 /dev/hidraw2

Reported-by: Enrik Berkhan <Enrik.Berkhan@inka.de>
Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: improve i2c large reads performance

The patch increases the read buffer size to 180 bytes. It reduces
the number of ft260_i2c_read() calls by three, improving the big
reads performance.

HID: ft260: improve i2c large reads performance

The patch increases the read buffer size to 180 bytes. It reduces
the number of ft260_i2c_read() calls by three, improving the big
reads performance.

$ sudo i2ctransfer -y -f 13 w2@0x51 0x0 0x0 r180

Before:

[ +4.071878] ft260_i2c_write_read: off 0x0 rlen 180 wlen 2
[ +0.000005] ft260_i2c_write: rep 0xd0 addr 0x51 off 0 len 2 wlen 2 flag 0x2 d[0] 0x0
[ +0.001097] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000175] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000004] ft260_i2c_read: rep 0xc2 addr 0x51 len 180 rlen 60 flag 0x3
[ +0.008579] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.000208] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000001] ft260_i2c_read: rep 0xc2 addr 0x51 len 120 rlen 60 flag 0x0
[ +0.008794] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.000181] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000002] ft260_i2c_read: rep 0xc2 addr 0x51 len 60 rlen 60 flag 0x4
[ +0.008817] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.000223] ft260_xfer_status: bus_status 0x20, clock 100

After:

[ +11.611642] ft260_i2c_write_read: off 0x0 rlen 180 wlen 2
[ +0.000005] ft260_i2c_write: rep 0xd0 addr 0x51 off 0 len 2 wlen 2 flag 0x2 d[0] 0x0
[ +0.008001] ft260_xfer_status: bus_status 0x20, clock 100
[ +0.000001] ft260_i2c_read: rep 0xc2 addr 0x51 len 180 rlen 180 flag 0x7
[ +0.008994] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.007987] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.007992] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.000206] ft260_xfer_status: bus_status 0x20, clock 100

Suggested-by: Enrik Berkhan <Enrik.Berkhan@inka.de>
Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: support i2c reads greater than HID report size

A random i2c read operation in EEPROM devices is implemented as a dummy
write operation, followed by a current address read operation. The

HID: ft260: support i2c reads greater than HID report size

A random i2c read operation in EEPROM devices is implemented as a dummy
write operation, followed by a current address read operation. The dummy
write operation is used to load the target byte or word address (a.k.a
offset) into the offset counter, from which the subsequent read operation
then reads.

To support longer than one HID report size random read, the ft260 driver
issues multiple pairs of i2c write offset + read data transactions of HID
report size so that the EEPROM device sees many i2c random read requests
from different offsets.

Two issues with the current implementation:
- This approach suffers from extra overhead caused by writing offset
requests.
- Necessity to handle offset per HID report in big-endian representation
as EEPROM devices expect. The current implementation does not do it and
correctly handles the reads up to 60 bytes only.

This patch addresses both issues by implementing more efficient approach.
It issues a single i2c read request of up to the EEPROM page size and then
waits for the data to arrive in multiple HID reports. For example, to read
the 256 bytes from a 24LC512 chip, which has 128 bytes page size, the old
method performs six ft260_i2c_write_read transactions while the new - two
only.

Before:

$ sudo ./i2cperf -d 2 -o 2 -s 128 -r 0-0xff 13 0x51 -S

Read block via i2ctransfer by chunks
-------------------------------------------------------------------
data rate(bps) efficiency(%) data size(B) total IOs IO size(B)
-------------------------------------------------------------------
40803 85 256 2 128

Kernel log of a single 128 bytes read request:

[ +2.376308] ft260_i2c_write_read: read_off 0x0 left_len 128 len 60
[ +0.000002] ft260_i2c_write: rep 0xd0 addr 0x51 off 0 len 2 wlen 2 flag 0x2 d[0] 0x0
[ +0.000707] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000173] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000001] ft260_i2c_read: rep 0xc2 addr 0x51 len 60
[ +0.008660] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.000156] ft260_xfer_status: bus_status 0x20, clock 100
[ +0.000001] ft260_i2c_write_read: read_off 0x3c left_len 68 len 60
[ +0.000001] ft260_i2c_write: rep 0xd0 addr 0x51 off 0 len 2 wlen 2 flag 0x2 d[0] 0x3c
[ +0.001034] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000191] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000001] ft260_i2c_read: rep 0xc2 addr 0x51 len 60
[ +0.008614] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.000203] ft260_xfer_status: bus_status 0x20, clock 100
[ +0.000001] ft260_i2c_write_read: read_off 0x78 left_len 8 len 8
[ +0.000001] ft260_i2c_write: rep 0xd0 addr 0x51 off 0 len 2 wlen 2 flag 0x2 d[0] 0x78
[ +0.000987] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000192] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000001] ft260_i2c_read: rep 0xc2 addr 0x51 len 8
[ +0.002614] ft260_raw_event: i2c resp: rep 0xd1 len 8
[ +0.000200] ft260_xfer_status: bus_status 0x20, clock 100

After:

$ sudo ./i2cperf -d 2 -o 2 -s 128 -r 0-0xff 13 0x51 -S

Read block via i2ctransfer by chunks
-------------------------------------------------------------------
data rate(bps) efficiency(%) data size(B) total IOs IO size(B)
-------------------------------------------------------------------
43990 85 256 2 128

Kernel log of a single 128 bytes read request:

[ +1.464346] ft260_i2c_write_read: off 0x0 rlen 128 wlen 2
[ +0.000002] ft260_i2c_write: rep 0xd0 addr 0x51 off 0 len 2 wlen 2 flag 0x2 d[0] 0x0
[ +0.001653] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000188] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000002] ft260_i2c_read: rep 0xc2 addr 0x51 len 128 rlen 60 flag 0x3
[ +0.008609] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.000157] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000002] ft260_i2c_read: rep 0xc2 addr 0x51 len 68 rlen 60 flag 0x0
[ +0.008840] ft260_raw_event: i2c resp: rep 0xde len 60
[ +0.000203] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000002] ft260_i2c_read: rep 0xc2 addr 0x51 len 8 rlen 8 flag 0x4
[ +0.002794] ft260_raw_event: i2c resp: rep 0xd1 len 8
[ +0.000201] ft260_xfer_status: bus_status 0x20, clock 100

Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Tested-by: Guillaume Champagne <champagne.guillaume.c@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: support i2c writes larger than HID report size

To support longer than one HID report size write, the driver splits a
single i2c message data payload into multiple i2c messages of HID rep

HID: ft260: support i2c writes larger than HID report size

To support longer than one HID report size write, the driver splits a
single i2c message data payload into multiple i2c messages of HID report
size. However, it does not replicate the offset bytes within the EEPROM
chip in every consequent HID report because it is not and should not be
aware of the EEPROM type. It breaks the i2c write message integrity and
causes the EEPROM device not to acknowledge the second HID report keeping
the i2c bus busy until the ft260 controller reports failure.

This patch preserves the i2c write message integrity by manipulating the
i2c flag bits across multiple HID reports to be seen by the EEPROM device
as a single i2c write transfer.

Before:

$ sudo ./i2cperf -f 2 -o 2 -s 64 -r 0-0xff 13 0x51 -S
Error: Sending messages failed: Input/output error

[ +3.667741] ft260_i2c_write: rep 0xde addr 0x51 off 0 len 60 d[0] 0x0
[ +0.007330] ft260_hid_output_report_check_status: wait 6400 usec, len 64
[ +0.000203] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.000001] ft260_i2c_write: rep 0xd1 addr 0x51 off 60 len 6 d[0] 0x0
[ +0.002337] ft260_hid_output_report_check_status: wait 1000 usec, len 10
[ +0.000157] ft260_xfer_status: bus_status 0x2e, clock 100
[ +0.000241] ft260_i2c_reset: done
[ +0.000003] ft260_i2c_write: failed to start transfer, ret -5

After:

$ sudo ./i2cperf -f 2 -o 2 -s 128 -r 0-0xff 13 0x51 -S

Fill block with increment via i2ctransfer by chunks
-------------------------------------------------------------------
data rate(bps) efficiency(%) data size(B) total IOs IO size(B)
-------------------------------------------------------------------
71260 86 256 2 128

Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Tested-by: Guillaume Champagne <champagne.guillaume.c@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: improve i2c write performance

The patch improves the I2C write performance by 20 - 30 percent by
revising the sleep time in the ft260_hid_output_report_check_status()
in the following wa

HID: ft260: improve i2c write performance

The patch improves the I2C write performance by 20 - 30 percent by
revising the sleep time in the ft260_hid_output_report_check_status()
in the following ways:

1. Reduce the wait time and start to poll earlier.

Sending a large amount of data at a low I2C clock rate saturates the
internal FT260 buffer and causes hiccups in status readiness, as shown
below in the log fragment. Aligning the status check wait time to the
worst case significantly reduces the write performance.

[Oct22 10:28] ft260_i2c_write: rep 0xd8 addr 0x51 off 0 len 34 d[0] 0x0
[ +0.005296] ft260_xfer_status: bus_status 0x20, clock 100
[ +0.013460] ft260_i2c_write: rep 0xd8 addr 0x51 off 0 len 34 d[0] 0x0
[ +0.003244] ft260_hid_output_report_check_status: wait 1920 usec, len 38
[ +0.000190] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.015324] ft260_i2c_write: rep 0xd8 addr 0x51 off 0 len 34 d[0] 0x0
[ +0.003491] ft260_hid_output_report_check_status: wait 1920 usec, len 38
[ +0.000202] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.016047] ft260_i2c_write: rep 0xd8 addr 0x51 off 0 len 34 d[0] 0x0
[ +0.002768] ft260_hid_output_report_check_status: wait 1920 usec, len 38
[ +0.000150] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.011389] ft260_i2c_write: rep 0xd8 addr 0x51 off 0 len 34 d[0] 0x0
[ +0.003467] ft260_hid_output_report_check_status: wait 1920 usec, len 38
[ +0.000191] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000172] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000131] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000241] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000233] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000190] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000196] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.011314] ft260_i2c_write: rep 0xd8 addr 0x51 off 0 len 34 d[0] 0x0
[ +0.003334] ft260_hid_output_report_check_status: wait 1920 usec, len 38
[ +0.000227] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000204] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000198] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000147] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.011060] ft260_i2c_write: rep 0xd8 addr 0x51 off 0 len 34 d[0] 0x0

Before:
$ sudo ./i2cperf -f 2 -o 2 -s 32 -r 0-0xff 13 0x51 -S

Fill block with increment via i2ctransfer by chunks
-------------------------------------------------------------------
data rate(bps) efficiency(%) data size(B) total IOs IO size(B)
-------------------------------------------------------------------
40510 80 256 8 32

After:
$ sudo ./i2cperf -f 2 -o 2 -s 32 -r 0-0xff 13 0x51 -S

Fill block with increment via i2ctransfer by chunks
-------------------------------------------------------------------
data rate(bps) efficiency(%) data size(B) total IOs IO size(B)
-------------------------------------------------------------------
52584 80 256 8 32

2. Do not sleep if the estimated I2C transfer time is below 2 ms since
the first xfer status query frequently takes around 1.5 ms, and the
following status queries take about 200us on average. So we usually
return from the routine after the first 1 - 3 status checks.

[Oct22 11:14] ft260_i2c_write: rep 0xd4 addr 0x51 off 0 len 18 d[0] 0x0
[ +0.004270] ft260_xfer_status: bus_status 0x20, clock 100
[ +0.013889] ft260_i2c_write: rep 0xd4 addr 0x51 off 0 len 18 d[0] 0x0
[ +0.000856] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000138] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.013352] ft260_i2c_write: rep 0xd4 addr 0x51 off 0 len 18 d[0] 0x0
[ +0.001501] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000177] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.014477] ft260_i2c_write: rep 0xd4 addr 0x51 off 0 len 18 d[0] 0x0
[ +0.001377] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000233] ft260_xfer_status: bus_status 0x41, clock 100
[ +0.000191] ft260_xfer_status: bus_status 0x40, clock 100
[ +0.013197] ft260_i2c_write: rep 0xd4 addr 0x51 off 0 len 18 d[0] 0x0

Before:
$ sudo ./i2cperf -f 2 -o 2 -s 16 -r 0-0xff 13 0x51 -S

Fill block with increment via i2ctransfer by chunks
-------------------------------------------------------------------
data rate(bps) efficiency(%) data size(B) total IOs IO size(B)
-------------------------------------------------------------------
28826 73 256 16 16

After:
$ sudo ./i2cperf -f 2 -o 2 -s 16 -r 0-0xff 13 0x51 -S

Fill block with increment via i2ctransfer by chunks
-------------------------------------------------------------------
data rate(bps) efficiency(%) data size(B) total IOs IO size(B)
-------------------------------------------------------------------
45138 73 256 16 16

Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Tested-by: Guillaume Champagne <champagne.guillaume.c@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

HID: ft260: ft260_xfer_status routine cleanup

After clarifying with FTDI's support, it turned out that the error
condition (bit 1) in byte 1 of the i2c status HID report is a status
bit reflecting a

HID: ft260: ft260_xfer_status routine cleanup

After clarifying with FTDI's support, it turned out that the error
condition (bit 1) in byte 1 of the i2c status HID report is a status
bit reflecting all error conditions. When bits 2, 3, or 4 are raised
to 1, bit 1 is set to 1 also. Since the ft260_xfer_status routine tests
the error condition bit and exits in the case of an error, the program
flow never reaches the conditional expressions for 2, 3, and 4 bits when
any of them indicates an error state. Though these expressions are never
evaluated to true, they are checked several times per IO, increasing the
ft260_xfer_status polling cycle duration.

The patch removes the conditional expressions for 2, 3, and 4 bits in
byte 1 of the i2c status HID report.

Signed-off-by: Michael Zaidman <michael.zaidman@gmail.com>
Tested-by: Guillaume Champagne <champagne.guillaume.c@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

show more ...

Merge tag 'v5.18'

Linux 5.18