418bba1706
Constant acceleration detector now blocks mouse movement well before
...
causing a full lockout
7 years ago
f1bb58e8c2
Maintain constant acceleration counter for several seconds, on the
...
assumption that a mouse bot would wait until the human stops using the
mouse before starting to do anything bad.
7 years ago
d7f66cd393
Better support for multi-interface HID: Logitech unifying receivers
...
- Select mouse interface first
- Correctly specify active interface when retrieving HID descriptors
7 years ago
c55b525c41
Support HID mouse reports with an ID header (Lenovo)
7 years ago
8403c9ee93
Restore jiggle detection.
...
This is no longer taken care of by the constant acceleration detector,
so we need to do it explicitly
7 years ago
d6f66e67bd
Reset mouse velocity history on each movement
7 years ago
3d91aa2115
Implement random-movement credit in the constant acceleration detector.
...
This makes it more forgiving of constant acceleration mid-movement, if
the beginning of the movement looks random. Handy for mice that for some
reason generate bot-like constant velocities.
7 years ago
7e4af61910
Tweak keyboard lockout threshold
7 years ago
7638e39f67
Mouse botdetect config comments
7 years ago
0c3c3a033c
Remove jiggle detector as the constant-acceleration detector already
...
does the same job.
7 years ago
84bd03b539
Implemented jiggle detection
7 years ago
8c6d54d110
Tweaks to mouse jump detection
7 years ago
b052645faa
Mouse jump detection implmented
7 years ago
324f115476
Implement mouse acceleration pulse timing. Doesn't work on realistic
...
inputs.
7 years ago
0baf8ea52b
Extend 'fast' key histogram out to 300ms
7 years ago
2deea5a646
Added error band comparison to velocity matching
7 years ago
f7b7ce84e7
Initial implementation of mouse bot detection
7 years ago
302bca3690
Re-implement keyboard bot detection based on timing histograms, rather
...
than maximum typing speed.
7 years ago
441caf6768
Tweaks to HID keyboard bot detection
7 years ago
fa00fd95dd
HID BotDetect functionality debugged and working:
...
- Implement advanced LED flash support
- Tweak HAL_SPI_TransmitReceive to tolerate a longer SysTick interrupt
- Tune KEYBOARD_BOTDETECT parameters
- L-shift and R-shift are now high-speed alphanumeric keys
7 years ago
02f6842723
HID keyboard bot detection mostly implemented. Needs more advanced LED
...
flash code, and testing.
7 years ago
fc7d0b67d7
Enable all config options by default
7 years ago
cf1d5da8e8
Build options complete & tested:
...
- Implement SCSI WP flag in mode parameter header
- Refactor #ifdef names
- #ifdef out more HID code
8 years ago
2cd86a7201
Tweak write protect SCSI sense code
8 years ago
3a94c2c863
Implemented optional build for mass storage, keyboard, and mouse
...
(Upstream). Also implemented read-only mass storage mode.
Todo:
- Test. Everything!!!
- Report write protect flag in SCSI "mode parameter header" (whatever
that is)
8 years ago
7cf9b67d1f
Implemented read-only option (Downstream only)
8 years ago
53fea530aa
Implemented optional build for mass storage, keyboard, and mouse
...
features.
8 years ago
7419c0f6a4
Create README.md
8 years ago
1b1567243e
Remove obsolete doc images
8 years ago
482acf6cee
Update README.md
8 years ago
b476df268a
Fixed devices-shrinking-by-one-block problem.
...
So it turns out the SCSI ReadCapacity10 command does not return the
device capacity. It returns the address of the last block, which is
(capacity - 1).
Also fixed Upstream's end-of-device sanity check.
8 years ago
e2404109a2
Fix for the previous fix. This time for sure.
...
Previously we wrote 4 x 64 byte transactions to the downstream device.
This is fast, but very occasionally the USB host silicon would bug out.
Reverting to single 64 byte write transactions increases reliability at
the expense of some write speed.
8 years ago
7022b16a71
Another fix for occasional mass storage write freakouts.
...
USB host controller sometimes incorrectly reports number of TX packets
in queue.
8 years ago
a287c39b84
Fix for occasional freakouts on mass storage writes.
...
Upstream's edge-detect interrupt that starts SPI transmission was not
firing. Sometimes. On some boards only. The cause was Downstream not
allowing enough time between edges for Upstream to catch the signal.
Also misc code tidyups.
8 years ago
d38dab22ff
Meaningless update to project settings file
8 years ago
6cabce2538
Update README.md
8 years ago
59714a7357
Update README.md
8 years ago
17eecf73c9
Merge branch 'USG_1.0' of https://github.com/robertfisk/usg into USG_1.0
8 years ago
8760a89f21
Remove obsolete kicad library from beta 1 project
8 years ago
ad5dbe5101
Update README.md
8 years ago
2843654ff7
Flush HID endpoint on init.
...
Fixes delayed HID events after failed wakeup attempt.
8 years ago
747681769b
Fix yesterday's HID disconnect/reconnect fix
8 years ago
eea91c8077
Fix for Upstream JTAG lockout - don't execute WFI too soon!
8 years ago
41c0624459
Startup LED flash is now asynchronous.
...
Not only a faster startup process, but also catches some HID devices
that only work if initialised within 500ms of powerup.
8 years ago
42d6030e7b
Refactor and improve HID report item parsing
8 years ago
0e42cf71dd
Correctly resume operation after HID device unplug/replug
8 years ago
a98af92c23
Correctly parse HID report descriptor items with unusual lengths.
...
Thanks Logitech.
8 years ago
e63525f868
Don't freakout when wakeup-from-sleep fails during keyboard report out.
8 years ago
7a1fe0fb65
Fix for mouse movement before waking from suspend
8 years ago
46ecbd410d
After an epic battle, suspend/resume/wakeup is now supported.
...
It turns out that suspend support on the STM32 USB core is buggy as
heck. Host mode cannot resume after suspend, and device mode cannot
receive resume or send wakeup signalling.
I managed to fake resume support by keeping Downstream and our connected
device running at full power, and simulating a wakeup event to the host
by disconnecting/reconnecting Upstream from the host.
8 years ago
Robert Fisk