[Bluez-users] hcid dying

[Bluez-users] hcid dying

[Edd Dumbill]

Recently I've experienced the hcid process dying without warning.  I've
not yet been able to isolate why this happens, but another user reported
the same thing happening to me today, so I thought I'd ask if anybody
else was seeing this happen.

-- Edd




-------------------------------------------------------
This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code NWMGYKND
_______________________________________________
Bluez-users mailing list
Bluez-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-users

Re: [Bluez-users] hcid dying

[Marcel Holtmann]

Hi Edd,

> Recently I've experienced the hcid process dying without warning.  I've
> not yet been able to isolate why this happens, but another user reported
> the same thing happening to me today, so I thought I'd ask if anybody
> else was seeing this happen.

I don't saw that on my machines. What distribution? What kernel?

Regards

Marcel




-------------------------------------------------------
This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code NWMGYKND
_______________________________________________
Bluez-users mailing list
Bluez-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-users

Re: [Bluez-users] hcid dying

[Edd Dumbill]

On Mon, 2004-06-21 at 13:44 +0200, Marcel Holtmann wrote:
> Hi Edd,
> 
> > Recently I've experienced the hcid process dying without warning.  I've
> > not yet been able to isolate why this happens, but another user reported
> > the same thing happening to me today, so I thought I'd ask if anybody
> > else was seeing this happen.
> 
> I don't saw that on my machines. What distribution? What kernel?

Debian (so I'm not excluding that this might be an error I introduced
somehow :).

For me, 2.6.6-mh3 when it last happened, iirc.

For the bug reporter I spoke to, 2.4.26-1-686 (from Debian.)

I asked for diagnostics from the reporter, and he sent me the output of
hciconfig -a, just after he restarted hcid:

hci0:   Type: USB
        BD Address: 00:0A:9A:xx:xx:xx ACL MTU: 339:4  SCO MTU: 64:0
        UP RUNNING PSCAN ISCAN
        RX bytes:1978 acl:60 sco:0 events:53 errors:0
        TX bytes:1111 acl:32 sco:0 commands:17 errors:0
        Features: 0xff 0xff 0x3d 0x00 0x00 0x00 0x00 0x00
        Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
        Link policy: RSWITCH HOLD SNIFF PARK
        Link mode: SLAVE ACCEPT
        Name: 'hactar-0202CC89E0E83C4EC7A99FE7ED3D12E970B...2.c'
        Class: 0x100100
        Service Classes: Object Transfer
        Device Class: Computer, Uncategorized
        HCI Ver: 1.1 (0x1) HCI Rev: 0x93 LMP Ver: 1.1 (0x1) LMP Subver:
0x93
        Manufacturer: Transilica, Inc. (24)

notice the weird name.  This is with hcid.conf saying:

	name "%h-%d";

which is the default for Debian.

Looks like there might be some bug in computing the device number.  It
should be reading "hactar-0".

-- Edd

Re: [Bluez-users] hcid dying

[Marcel Holtmann]

[-- Attachment #1: Type: text/plain, Size: 1590 bytes --]

Hi Edd,

> Debian (so I'm not excluding that this might be an error I introduced
> somehow :).
> 
> For me, 2.6.6-mh3 when it last happened, iirc.
> 
> For the bug reporter I spoke to, 2.4.26-1-686 (from Debian.)
> 
> I asked for diagnostics from the reporter, and he sent me the output of
> hciconfig -a, just after he restarted hcid:
> 
> hci0:   Type: USB
>         BD Address: 00:0A:9A:xx:xx:xx ACL MTU: 339:4  SCO MTU: 64:0
>         UP RUNNING PSCAN ISCAN
>         RX bytes:1978 acl:60 sco:0 events:53 errors:0
>         TX bytes:1111 acl:32 sco:0 commands:17 errors:0
>         Features: 0xff 0xff 0x3d 0x00 0x00 0x00 0x00 0x00
>         Packet type: DM1 DM3 DM5 DH1 DH3 DH5 HV1 HV2 HV3
>         Link policy: RSWITCH HOLD SNIFF PARK
>         Link mode: SLAVE ACCEPT
>         Name: 'hactar-0202CC89E0E83C4EC7A99FE7ED3D12E970B...2.c'
>         Class: 0x100100
>         Service Classes: Object Transfer
>         Device Class: Computer, Uncategorized
>         HCI Ver: 1.1 (0x1) HCI Rev: 0x93 LMP Ver: 1.1 (0x1) LMP Subver:
> 0x93
>         Manufacturer: Transilica, Inc. (24)
> 
> notice the weird name.  This is with hcid.conf saying:
> 
> 	name "%h-%d";
> 
> which is the default for Debian.
> 
> Looks like there might be some bug in computing the device number.  It
> should be reading "hactar-0".

the device number is fine, but the expand_name() function is maybe
wrong. I don't know if this related somehow, but if it overwrites the
stack everything can happen. Here is a patch that restricts the device
name. It is untested so I don't know if it works.

Regards

Marcel


[-- Attachment #2: patch --]
[-- Type: text/x-patch, Size: 1939 bytes --]

Index: hcid/lib.c
===================================================================
RCS file: /cvsroot/bluez/utils/hcid/lib.c,v
retrieving revision 1.4
diff -u -b -w -B -r1.4 lib.c
--- hcid/lib.c	28 Apr 2004 12:09:32 -0000	1.4
+++ hcid/lib.c	21 Jun 2004 12:28:22 -0000
@@ -53,7 +53,7 @@
  * Device name expansion 
  * 	%d - device id
  */
-char *expand_name(char *dst, char *str, int dev_id)
+char *expand_name(char *dst, int size, char *str, int dev_id)
 {
 	register int sp, np, olen;
 	char *opt, buf[10];
@@ -62,7 +62,7 @@
 		return NULL;
 
 	sp = np = 0;
-	while (str[sp]) {
+	while (np < size - 1 && str[sp]) {
 		switch (str[sp]) {
 		case '%':
 			opt = NULL;
@@ -88,6 +88,7 @@
 			if (opt) {
 				/* substitute */
 				olen = strlen(opt);
+				if (np + olen < size - 1)
 				memcpy(dst + np, opt, olen);
 				np += olen;
 			}
Index: hcid/lib.h
===================================================================
RCS file: /cvsroot/bluez/utils/hcid/lib.h,v
retrieving revision 1.3
diff -u -b -w -B -r1.3 lib.h
--- hcid/lib.h	28 Apr 2004 12:09:32 -0000	1.3
+++ hcid/lib.h	21 Jun 2004 12:28:22 -0000
@@ -30,7 +30,7 @@
 
 #include <errno.h>
 
-char *expand_name(char *dst, char *str, int dev_id);
+char *expand_name(char *dst, int size, char *str, int dev_id);
 
 char *get_host_name(void);
 
Index: hcid/main.c
===================================================================
RCS file: /cvsroot/bluez/utils/hcid/main.c,v
retrieving revision 1.15
diff -u -b -w -B -r1.15 main.c
--- hcid/main.c	7 May 2004 23:08:03 -0000	1.15
+++ hcid/main.c	21 Jun 2004 12:28:22 -0000
@@ -227,7 +227,8 @@
 	/* Set device name */
 	if (device_opts->name) {
 		change_local_name_cp cp;
-		expand_name(cp.name, device_opts->name, hdev);
+		memset(cp.name, 0, sizeof(cp.name));
+		expand_name(cp.name, sizeof(cp.name), device_opts->name, hdev);
 
 		hci_send_cmd(s, OGF_HOST_CTL, OCF_CHANGE_LOCAL_NAME,
 			CHANGE_LOCAL_NAME_CP_SIZE, (void *) &cp);