KDL on newer Apple trackpad/internal keyboard

[DHowett]

I'm pretty sure that something about the USB devices in newer apple laptops causes the kernel to explode.

Touch the trackpad/internal keyboard just outside the bootloader and it panics normally, it panics so hard that it won't even display the backtrace. I managed to get one though.

[jscipione]

KDL of crash

[jscipione]

This is coming from IRC, I am not the bug reporter so I don't have all the details.

[jscipione]

Check if name is NULL before calling strcmp() on it. This should prevent the KDL here by returning a NULL command instead

[pulkomandy]

This patch doesn't look right, for several reasons.

1) There is no reason for this command to be called with a NULL string. If it happens, it's better to crash so the actual bug can be fixed, instead of hiding the issue. 2) In the attached KDL, nothing points at the passed string being NULL. Parameters are passed on the stack, so you couldn't see them in that picture. But, if the problem was dereferencing a NULL pointer, the message would say "unhandled page fault at 0", and it says some other address. I think you saw the 0 in EAX, but that is probably the index into the strings used by strcmp. 3) If we check the code in the HID keyboard driver (src/add-ons/kernel/drivers/input/usb_hid/KeyboardProtocolHandler.cpp), we can see that the string passed is a constant, and cannot be NULL. But this also means, it cannot be the cause of the page fault if everything is going ok.

So, this issue points at some memory corruption iside the kernel. Maybe the string was overwritten with something, or maybe the list of KDL commands is corrupt.

Is the problem still valid? Can anyone reproduce it in recent versions of Haiku?

[waddlesplash]

No reply in 2 years.

[nielx]

Remove milestone for tickets with status = closed and resolution != fixed