Re: Regression from 2.6.21 to 2.6.22 and above

On Fri, 12 Oct 2007 15:52:51 -0700
Pete Zaitcev <[email protected]> wrote:

> On Fri, 12 Oct 2007 14:14:32 +0100, Alan Cox <[email protected]> wrote:
> 
> > > 6de02123bf3e8baeee97fff7efc50bc192332804 is first bad commit
> > > Author: Alan Cox <[email protected]>
> > > Date:   Tue May 8 00:26:21 2007 -0700
> > > 
> > >     tty: i386/x86_64 arbitary speed support
> 
> > > Hmm, that doesn't seem to be an usb issue at all. My guess is, that the
> > > provided libct.so can't handle the new termios struct. What do you think?
> 
> > Nothing should break as the termios structures and bits have not changed.
> > New ones have been added and translation done.
> 
> Actually, this bit is what corrupts memory in libct:
> 
> -#define kernel_termios_to_user_termios(u, k) copy_to_user(u, k, sizeof(struct termios))
> +#define kernel_termios_to_user_termios(u, k) copy_to_user(u, k, sizeof(struct termios2))
> 
> The problem is, Stefan uses kobil_sct, which is the last driver
> in the tree which does this:
> 
> 	case TCSETS:   // 0x5402
> 		if (user_termios_to_kernel_termios(&priv->internal_termios,
> 						   (struct ktermios __user *)arg))
> 			return -EFAULT;
> 
> That driver has to be given the same treatment as the rest of the
> serial drivers.

The patches for that have been with Greg for a bit and I saw the last
patch storm he just posted included them. Basically kobil was broken. Its
trying to intercept the termios support and you can't do that, its never
been right to do so.

Alan

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel