search

raspberrypi / picotool

BSD 3-Clause "New" or "Revised" License
587 stars 103 forks source link

picotool silently segfaults if it can't open a device #151

Open usedbytes opened 2 weeks ago

usedbytes commented 2 weeks ago

This is ~ a re-report of #119 , but I think the report there is unclear/misleading.

If picotool can't open the device (ret = libusb_open(device, dev_handle); fails), then handle will end up as 0x0 in the device map, and (for example) reboot_device will segfault.

My Pico (RP2040) is reporting vid:pid 2e8a:000a (Note the HW is actually a plain Pico, not a Badger2040, but that shouldn't matter)

$ lsusb
Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 001 Device 006: ID 2e8a:000a Pimoroni Badger2040
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
$ uname -a
Linux eddie 6.6.21-current-meson64 #1 SMP PREEMPT Wed Mar  6 14:48:45 UTC 2024 aarch64 aarch64 aarch64 GNU/Linux
$ ./picotool reboot -f
Segmentation fault
$ sudo ./picotool reboot -f
The device was asked to reboot into application mode.

At the very least I think the null handle should be checked/caught and something useful reported to the user ("Couldn't open device"), rather than segfaulting.

lurch commented 2 weeks ago

OffTopic: https://github.com/raspberrypi/usb-pid says 2e8a:000a is "Raspberry Pi Pico SDK CDC UART (RP2040)" so I've no idea why lsusb is listing it as "Pimoroni Badger2040" ??

usedbytes commented 2 weeks ago

That’s just a string descriptor reported by my FW. lsusb doesn’t have a database or anything, it’s just printing what the device says in the Manufacturer and Product ID strings.

I originally wrote the code for a Badger 2040 and looks like I never changed the string.

will-v-pi commented 6 days ago

I think I've found the cause of this segfault - could you test with that new branch and see if it fixes this for you? https://github.com/raspberrypi/picotool/tree/no-usb-segfault-fix