chadderz121 / csud Goto Github PK

had a quick glance over the readme and other issues, and dont see a clear answer

does this driver support both host and device modes? or only host mode?

i'm interested in using it to emulate a usb device on an rpi4, where the dwc controller is routed to the USB-C port

It can't input smoothly on my Raspberry b+1.2v I tried so many people's os for raspberry and It always Input a word then stuck :<
So I thought It could my keyboard problem.I want your keyboard and mouse model

you guys can help me @_@

Hey Alex,

I'm writing a micro kernel os, and trying to use your source code(https://github.com/MisaZhu/EwokOS/tree/master/system/platforms/raspix/drivers/usbd) . It works pretty well ,just wanna know do you have any plan to support more devices ?

Misa.Z

It has been reported that the KeyboardPoll method has the side effect of overwritting addresses 0 and 4. Have not tested this myself, said to occur if the keyboard has no activiy for a short time. Not always written as 0. Presumably due to writing to a NULL pointer.

I struggled many hours today to find out why the csud driver would initialize correctly but then fail to iterate the root hub. The problem was that since GCC4.4(?) the packed attribute has to be set on nested structs too if they need to be bit-packed. Example:

struct UsbConfigurationDescriptor {
    u8 DescriptorLength; // +0x0
    enum DescriptorType DescriptorType : 8; // +0x1
    u16 TotalLength; // +0x2
    u8 InterfaceCount; // +0x4
    u8 ConfigurationValue; // +0x5
    u8 StringIndex; // +0x6
    struct {
        unsigned _reserved0_4 : 5; // @0
        bool RemoteWakeup : 1; // @5
        bool SelfPowered : 1; // @6
        unsigned _reserved7 : 1; // @7
    } Attributes; // +0x7
    u8 MaximumPower; // +0x8
} __attribute__ ((__packed__));

sizeof(struct UsbConfigurationDescriptor) returns 12 for me, not 9 as the code implies. This breaks the descriptor discovery loop in UsbConfigure() because of wrong offsets.

Adding the packed attribute on the inner struct "Attributes" fixed the problem for me:

...
    struct {
        unsigned _reserved0_4 : 5; // @0
        bool RemoteWakeup : 1; // @5
        bool SelfPowered : 1; // @6
        unsigned _reserved7 : 1; // @7
    } __attribute__ ((__packed__)) Attributes; // +0x7
    u8 MaximumPower; // +0x8
} __attribute__ ((__packed__));

For reference, I am running GCC 4.7.4:
arm-none-eabi-gcc (GNU Tools for ARM Embedded Processors) 4.7.4 20130913 (release) [ARM/embedded-4_7-branch revision 202601]

Cheers,
Lukas

Got a few of these warnings when compiling it into my code based on BakingPi.

arm-none-eabi-ld --no-undefined build/_start.o build/mailbox.o build/printk.o build/fb.o b
uild/gpio.o build/font_8x8.o build/fb_test.o build/main.o build/led.o build/timer.o -Map k
ernel.map -L libs -lcsud -o build/output.elf -T kernel.ld
arm-none-eabi-ld: warning: libs/libcsud.a(device.keyboard.c.o) uses 2-byte wchar_t yet the
 output is to use 4-byte wchar_t; use of wchar_t values across objects may fail
arm-none-eabi-ld: Warning: libs/libcsud.a(device.keyboard.c.o): Unknown EABI object attrib
ute 44

I see -fshort-wchar is used. Any reason for that? Should I include it also in my makefile?

Thanks for this driver (and your tutorial!). (Hit ctrl-enter accidentally.. Sorry for the partial submit)

On MacOS, apostrophes and parens in the makefile intro text (shown by the 'all' target) cause errors in the shell. eg:

James-MacBook:csud jamescone$ make
/bin/sh: -c: line 0: unexpected EOF while looking for matching `''
/bin/sh: -c: line 1: syntax error: unexpected end of file
make: *** [all] Error 2
James-MacBook:csud jamescone$

I ahve a patch, but the markup mangles it. I've failed to work out how to attach it.

When I attempt to compile using "make driver config=FINAL type=LOWLEVEL target=RPI config=FINAL gnu=arm-none-eabi " I get the following:

make: *** No rule to make target 'build', needed by 'libcsud.a'. Stop.

Am I missing something? or is it just a misconfiguration?
Thanks in advance.

Debug log:

CSUD: USB driver version 1.0
CSUD: HID driver version 1.0
CSUD: Hub driver version 1.0
CSUD: Keyboard driver version 1.0
CSUD: Mouse driver version 1.0
HCD: Reserving memory.
Platform: First memory allocation, reserving 16KiB of heap, 256 entries.
Platform: malloc(0x400) = 0xa9db8. (1023/16384)
Platform: Out of memory! We should've had more allocations in platform.c.
Platform: Out of memory! We should've had more allocations in platform.c.
HCD: Hardware: OT2.80a (BCM2708a)
HCD: Internal DMA mode.
HCD: Hardware configuration: 00000000 228ddd50 0ff000e8 1ff00020
HCD: Host configuration: 00200000
HCD: Disabling interrupts.
HCD: Powering USB on.
HCD: Load completed.
HCD: Start core.
Platform: Out of memory! We should've had more allocations in platform.c.
USBD: Abort, HCD failed to start.
Platform: free(0xa9db8) (-1/16384)

source/device/hid/keyboard.c:163:7: warning: this 'if' clause does not guard... [-Wmisleading-indentation]

Are both lines supposed to be guarded by the if statement, or only the LOG_DEBUGF?