>Number:         15019
>Category:       bin
>Synopsis:       usbhidctl -l generates ugly error on device detach
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    bin-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Dec 21 00:18:00 PST 2001
>Closed-Date:
>Last-Modified:
>Originator:     David Sainty
>Release:        NetBSD 1.5X
>Organization:
Dynamic Technology Services and Products Ltd (NZ)
>Environment:
System: NetBSD tequila.dave.dtsp.co.nz 1.5X NetBSD 1.5X (TEQUILA) #0: Sat Dec 1 17:24:03 NZDT 2001 dave@tequila.dave.dtsp.co.nz:/vol/tequila/userD/NetBSD-tequila/src/sys/arch/i386/compile/TEQUILA i386
Architecture: i386
Machine: i386
>Description:
	When using the loop option to usbhidctl and the device is detached,
	usbhidctl generates a rather ugly error, like:

	usbhidctl: bad read -1 != 4: Input/Output error

	Given that it is easy to generate an error like this, the code should
	generate a cleaner error.  The code is currently a little careless, on
	the assumption that unexpected results would typically never happen.

>How-To-Repeat:
	Pull the plug on a USB HID device whilst looping with usbhidctl -l

>Fix:
--- src/usr.bin/usbhidctl/usbhid.c.orig	Tue Oct 23 19:27:20 2001
+++ src/usr.bin/usbhidctl/usbhid.c	Fri Dec 21 21:14:02 2001
@@ -570,9 +570,11 @@
 		ssize_t readlen;
 
 		readlen = read(hidfd, dbuf, dlen);
-		if (readlen < 0 || dlen != (size_t)readlen)
-			err(1, "bad read %ld != %ld",
-			    (long)readlen, (long)dlen);
+		if (readlen < 0)
+			err(1, "Device read error");
+		if (dlen != (size_t)readlen)
+			errx(1, "Unexpected response length: %lu != %lu",
+			     (unsigned long)readlen, (unsigned long)dlen);
 
 		collind = 0;
 		hdata = hid_start_parse(rd, 1 << hid_input);
>Release-Note:
>Audit-Trail:
>Unformatted: