Re: Patch to add usbmon

2005-02-04 Thread Marcel Holtmann
Hi Pete, > > While I am really thinking about starting usbdump, I may ask why you > > have choosen to use debugfs as interface. This will not be available in > > normal distribution kernels and I think a general USB monitoring ability > > would be great. For example like we have it for Ethernet,

Re: Patch to add usbmon

2005-02-04 Thread Greg KH
On Tue, Feb 01, 2005 at 10:25:13PM -0800, Pete Zaitcev wrote: > This patch adds so-called "usbmon", or USB monitoring framework, similar > to what tcpdump provides for Ethernet. This is an initial version, but > it should be safe and useful. It adds an overhead of an if () statement > into

Re: Patch to add usbmon

2005-02-04 Thread Greg KH
On Tue, Feb 01, 2005 at 10:25:13PM -0800, Pete Zaitcev wrote: This patch adds so-called usbmon, or USB monitoring framework, similar to what tcpdump provides for Ethernet. This is an initial version, but it should be safe and useful. It adds an overhead of an if () statement into submission

Re: Patch to add usbmon

2005-02-04 Thread Marcel Holtmann
Hi Pete, While I am really thinking about starting usbdump, I may ask why you have choosen to use debugfs as interface. This will not be available in normal distribution kernels and I think a general USB monitoring ability would be great. For example like we have it for Ethernet,

Re: Patch to add usbmon

2005-02-02 Thread Pete Zaitcev
On Wed, 02 Feb 2005 17:40:17 +0100, Marcel Holtmann <[EMAIL PROTECTED]> wrote: > While I am really thinking about starting usbdump, I may ask why you > have choosen to use debugfs as interface. This will not be available in > normal distribution kernels and I think a general USB monitoring

Re: Patch to add usbmon

2005-02-02 Thread Marcel Holtmann
Hi Pete, > > I think if cat is the prefered tool for viewing this file then it should > > be more human readable. If not, then a binary format should be choosen. > > Maybe we can implement both. Is this possible? > > Yes. Now you know why files were split as they were. still no reason for me to

Re: Patch to add usbmon

2005-02-02 Thread Marcel Holtmann
Hi Pete, I think if cat is the prefered tool for viewing this file then it should be more human readable. If not, then a binary format should be choosen. Maybe we can implement both. Is this possible? Yes. Now you know why files were split as they were. still no reason for me to split

Re: Patch to add usbmon

2005-02-02 Thread Pete Zaitcev
On Wed, 02 Feb 2005 17:40:17 +0100, Marcel Holtmann [EMAIL PROTECTED] wrote: While I am really thinking about starting usbdump, I may ask why you have choosen to use debugfs as interface. This will not be available in normal distribution kernels and I think a general USB monitoring ability

Re: Patch to add usbmon

2005-02-01 Thread Pete Zaitcev
This patch adds so-called "usbmon", or USB monitoring framework, similar to what tcpdump provides for Ethernet. This is an initial version, but it should be safe and useful. It adds an overhead of an if () statement into submission and giveback paths even when not monitoring, but this was deemed a

Re: Patch to add usbmon

2005-02-01 Thread Pete Zaitcev
On Tue, 01 Feb 2005 22:37:50 +0100, Marcel Holtmann <[EMAIL PROTECTED]> wrote: > I think if cat is the prefered tool for viewing this file then it should > be more human readable. If not, then a binary format should be choosen. > Maybe we can implement both. Is this possible? Yes. Now you know

Re: Patch to add usbmon

2005-02-01 Thread Marcel Holtmann
Hi Pete, > > By accident I removed the debugfs option from my kernel config and this > > makes usbmon totally useless. So I think the module approach is wrong > > from my point of view. Why not compile it always and if debugfs is > > available, then enable it when the usbcore gets loaded? > >

Re: Patch to add usbmon

2005-02-01 Thread Pete Zaitcev
On Tue, 01 Feb 2005 12:13:03 +0100, Marcel Holtmann <[EMAIL PROTECTED]> wrote: > By accident I removed the debugfs option from my kernel config and this > makes usbmon totally useless. So I think the module approach is wrong > from my point of view. Why not compile it always and if debugfs is >

Re: Patch to add usbmon

2005-02-01 Thread Marcel Holtmann
Hi Pete, today I just thought to give usbmon a try. Previously I used a hacked devio thing around usbfs_snoop to monitor the USB communication between a VMware and the Linux host. Greg, will such patch accepted for inclusion or will usbfs_snoop go away when usbmon is included? > > First off,

Re: Patch to add usbmon

2005-02-01 Thread Pete Zaitcev
On Mon, 31 Jan 2005 23:10:01 -0800, Greg KH <[EMAIL PROTECTED]> wrote: > First off, why make usbmon a module? You aren't allowing it to happen, > so just take out the parts of the patch that allow it. No, I do allow it. This way I can load and unload it when debugging it. Perhaps in the future

Re: Patch to add usbmon

2005-02-01 Thread Pete Zaitcev
On Mon, 31 Jan 2005 23:10:01 -0800, Greg KH [EMAIL PROTECTED] wrote: First off, why make usbmon a module? You aren't allowing it to happen, so just take out the parts of the patch that allow it. No, I do allow it. This way I can load and unload it when debugging it. Perhaps in the future we

Re: Patch to add usbmon

2005-02-01 Thread Marcel Holtmann
Hi Pete, today I just thought to give usbmon a try. Previously I used a hacked devio thing around usbfs_snoop to monitor the USB communication between a VMware and the Linux host. Greg, will such patch accepted for inclusion or will usbfs_snoop go away when usbmon is included? First off, why

Re: Patch to add usbmon

2005-02-01 Thread Pete Zaitcev
On Tue, 01 Feb 2005 12:13:03 +0100, Marcel Holtmann [EMAIL PROTECTED] wrote: By accident I removed the debugfs option from my kernel config and this makes usbmon totally useless. So I think the module approach is wrong from my point of view. Why not compile it always and if debugfs is

Re: Patch to add usbmon

2005-02-01 Thread Marcel Holtmann
Hi Pete, By accident I removed the debugfs option from my kernel config and this makes usbmon totally useless. So I think the module approach is wrong from my point of view. Why not compile it always and if debugfs is available, then enable it when the usbcore gets loaded? This may be

Re: Patch to add usbmon

2005-02-01 Thread Pete Zaitcev
On Tue, 01 Feb 2005 22:37:50 +0100, Marcel Holtmann [EMAIL PROTECTED] wrote: I think if cat is the prefered tool for viewing this file then it should be more human readable. If not, then a binary format should be choosen. Maybe we can implement both. Is this possible? Yes. Now you know why

Re: Patch to add usbmon

2005-02-01 Thread Pete Zaitcev
This patch adds so-called usbmon, or USB monitoring framework, similar to what tcpdump provides for Ethernet. This is an initial version, but it should be safe and useful. It adds an overhead of an if () statement into submission and giveback paths even when not monitoring, but this was deemed a

Re: Patch to add usbmon

2005-01-31 Thread Greg KH
On Mon, Jan 31, 2005 at 09:29:03PM -0800, Pete Zaitcev wrote: > This patch adds so-called "usbmon", or USB monitoring framework, similar > to what tcpdump provides for Ethernet. This is an initial version, but > it should be safe and useful. It adds an overhead of an if () statement > into

Patch to add usbmon

2005-01-31 Thread Pete Zaitcev
This patch adds so-called "usbmon", or USB monitoring framework, similar to what tcpdump provides for Ethernet. This is an initial version, but it should be safe and useful. It adds an overhead of an if () statement into submission and giveback paths even when not monitoring, but this was deemed a

Patch to add usbmon

2005-01-31 Thread Pete Zaitcev
This patch adds so-called usbmon, or USB monitoring framework, similar to what tcpdump provides for Ethernet. This is an initial version, but it should be safe and useful. It adds an overhead of an if () statement into submission and giveback paths even when not monitoring, but this was deemed a

Re: Patch to add usbmon

2005-01-31 Thread Greg KH
On Mon, Jan 31, 2005 at 09:29:03PM -0800, Pete Zaitcev wrote: This patch adds so-called usbmon, or USB monitoring framework, similar to what tcpdump provides for Ethernet. This is an initial version, but it should be safe and useful. It adds an overhead of an if () statement into submission