Re: Reboot resistant USB bug

[manu%netbsd.org@localhost (Emmanuel Dreyfus)]

Nick Hudson <nick.hudson%gmx.co.uk@localhost> wrote:

> Doesn't the hub power the port down? The bug is probably in uhub.c

I found a description of the problem in a chip maker technote:
https://www.nxp.com/docs/en/nxp/supporting-information/163020.pdf
"With the current PN533, it is possible to corrupt the USB Descriptors
information."

The workaround seems to exist in libnfc, in this file:
https://github.com/nfc-tools/libnfc/blob/master/libnfc/drivers/pn53x_usb.c

We find stuff like this:
  // Abuse the overflow bug to restore USB descriptors in one go
  log_put(LOG_GROUP, LOG_CATEGORY, NFC_LOG_PRIORITY_INFO, "%s", "Fixing USB
descriptors corruption");

But I suspect the way our kernel USB stack behaves prevent this from
being actually used.


-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu%netbsd.org@localhost