Skip to contentSkip to conversation

No more ptys!

Bluez-devel GUILLON Gabriel 3 participants 5 messages Aug 31, 2004 – Sep 14, 2004 1 message not in the archive Bluetooth drivers

lore
Share & tools
lore

GUILLON Gabriel

author
#0 lore raw

I'm playing around with 2 bluetooth dongle, under linux (me:Debian and other:Mandrake), trying rfcomm, hciattach, things like that. I do that using 'screen' commands.
After a will, I cannot create more screen: "no more pty", says screen!
I user kernel 2.6.8.1, bluez-utils 2.9-3
Anyone has a clue?

Signature · 18 lines
"Everything is funny, if you wait long enough"

Gab

"Vienne la nuit sonne l'heure
Les jours s'en vont je demeure."
- Guillaume Apollinaire - Le Pont Mirabeau -


-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click
_______________________________________________
Bluez-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-devel
maintainer
#1 lore raw

Hi Gabriel,

Quoted 5 lines

I'm playing around with 2 bluetooth dongle, under linux (me:Debian and other:Mandrake), trying rfcomm, hciattach, things like that. I do that using 'screen' commands.
After a will, I cannot create more screen: "no more pty", says screen! I user kernel 2.6.8.1, bluez-utils 2.9-3

this is actually not a Bluetooth problem.

Regards

Marcel

-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click
_______________________________________________
Bluez-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-devel

A message that is not in the archive (sent off-list or not yet received): [email protected]

maintainer
#3 lore raw

Hi Gabriel,

Quoted 9 lines
Quoted 8 lines
Quoted 5 lines

I'm playing around with 2 bluetooth dongle, under linux (me:Debian and other:Mandrake), trying rfcomm, hciattach, things like that. I do that using 'screen' commands.
After a will, I cannot create more screen: "no more pty", says screen! I user kernel 2.6.8.1, bluez-utils 2.9-3

this is actually not a Bluetooth problem.

Quoted 13 lines

This only append when I play with BT. It never appended before. Few hours before, I played with about 10 screen session with no problem. Then I reduced the number of screen sessions, played with BT, tried to open one more screen, then it failed!
I looked at /dev/pts: there were the number of opened screen session, no more, no less.
I closes everything but X, then it worked again.
I conclued with a memory problem (Mozilla, Thunderbird and OOo where open, and my computer is a small computer)
Day after, played again with BT, with not that much things opened (ie:
no ram problems, a priori)
Same problem arrived: no more ttys.
Time to reboot...

check the kernel mailing list, because this can't be a Bluetooth problem. There was a memory related to PTYs so you may wanna check out a newer kernel.

Regards

Marcel

-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click
_______________________________________________
Bluez-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-devel

#4 lore raw

freeing skbuff in l2cap_chan_send

In l2cap_chan_send (l2cap.c), it appears as though the initial skbuff allocated to store data
passed in the msg parameter is not freed
if the buffer must be fragmented to send
to HCI.

The skbuff and its associated fragment list
is passed to hci by calling hci_send_acl.

This in turn places all the individual fragments
on a tx queue.

The individual fragment skbuffs are freed when they are written to the tty device (I am looking at the case where HCI uart is used).

However, the original l2cap skbuff allocated to store the entire l2cap frame never appears to be freed.

I am looking at mh17 on a 2.4.20 kernel.

Is this a bug or am I missing something here?

Thanks.

Aaron

Signature · 10 lines
------------
Aaron Klish
Bluetooth Software Engineer
Motorola PCS
PH# (217) 384-8598
FX# (217) 384-8550

[X]  Motorola General Business Information
[ ]  Motorola Internal Use Only
[ ]  Motorola Confidential Proprietary
#5 lore raw

Re: [Bluez-devel] freeing skbuff in l2cap_chan_send

I was mistaken. This is not a bug.

Aaron

Quoted 27 lines · Aaron Klish

In l2cap_chan_send (l2cap.c), it appears as though the initial skbuff allocated to store data
passed in the msg parameter is not freed
if the buffer must be fragmented to send
to HCI.

The skbuff and its associated fragment list
is passed to hci by calling hci_send_acl.

This in turn places all the individual fragments
on a tx queue.

The individual fragment skbuffs are freed when they are written to the tty device (I am looking at the case where HCI uart is used).

However, the original l2cap skbuff allocated to store the entire l2cap frame never appears to be freed.

I am looking at mh17 on a 2.4.20 kernel.

Is this a bug or am I missing something here?

Thanks.

Aaron

Signature · 10 lines
------------
Aaron Klish
Bluetooth Software Engineer
Motorola PCS
PH# (217) 384-8598
FX# (217) 384-8550

[X]  Motorola General Business Information
[ ]  Motorola Internal Use Only
[ ]  Motorola Confidential Proprietary