Re: Perform mmap and poll on PUD character devices

2011-09-12 Thread Jean-Yves Migeon
On Sun, 11 Sep 2011 20:53:55 -0400, Thor Lancelot Simon wrote: On Sun, Sep 11, 2011 at 11:42:33PM +0200, Jean-Yves Migeon wrote: On 11.09.2011 21:07, Thor Lancelot Simon wrote: Similarly, I am not sure I believe the security justification, mostly because I don't really see why I should

Re: Perform mmap and poll on PUD character devices

2011-09-11 Thread Roger Pau Monné
I can't see how a userland implementation can have better performances than an all in-kernel implementation. There are more context switches, which are notr so cheap with Xen. From what I understand using blktap with Xen provides better performance because it is designed specifically for this

Re: Perform mmap and poll on PUD character devices

2011-09-11 Thread Manuel Bouyer
On Sun, Sep 11, 2011 at 08:58:53PM +0200, Roger Pau Monné wrote: I can't see how a userland implementation can have better performances than an all in-kernel implementation. There are more context switches, which are notr so cheap with Xen. From what I understand using blktap with Xen

Re: Perform mmap and poll on PUD character devices

2011-09-11 Thread Thor Lancelot Simon
On Sun, Sep 11, 2011 at 08:58:53PM +0200, Roger Pau Monn? wrote: I can't see how a userland implementation can have better performances than an all in-kernel implementation. There are more context switches, which are notr so cheap with Xen. From what I understand using blktap with Xen

Re: Perform mmap and poll on PUD character devices

2011-09-11 Thread Jean-Yves Migeon
On 11.09.2011 21:07, Thor Lancelot Simon wrote: Similarly, I am not sure I believe the security justification, mostly because I don't really see why I should believe that the very complex memory management code involved in providing the split kernel/user memory-mapped device driver interface

Re: Perform mmap and poll on PUD character devices

2011-09-11 Thread Thor Lancelot Simon
On Sun, Sep 11, 2011 at 11:42:33PM +0200, Jean-Yves Migeon wrote: On 11.09.2011 21:07, Thor Lancelot Simon wrote: Similarly, I am not sure I believe the security justification, mostly because I don't really see why I should believe that the very complex memory management code involved in

Re: Perform mmap and poll on PUD character devices

2011-09-10 Thread Roger Pau Monné
Hello, Thanks for the reply, sometimes I'm so focused in the problem that I forget the big picture. PUD is a framework present in NetBSD that allows to implement character and block devices in userspace. [1] The blktap (“block tap”) userspace toolkit provides a user-level disk I/O interface in

Re: Perform mmap and poll on PUD character devices

2011-09-10 Thread Thor Lancelot Simon
On Sat, Sep 10, 2011 at 12:05:53PM +0200, Roger Pau Monn? wrote: Hello, Thanks for the reply, sometimes I'm so focused in the problem that I forget the big picture. PUD is a framework present in NetBSD that allows to implement character and block devices in userspace. [1] The blktap

Re: Perform mmap and poll on PUD character devices

2011-09-10 Thread Mouse
I do not understahd why it is desirable to involve additional context switches to and from userspace into this data path. Instead of writing a bunch of fairly dubious page mapping code [...] in the kernel to support user-space daemons handling various virtual disk formats, why not put the

Re: Perform mmap and poll on PUD character devices

2011-09-08 Thread Erik Fair
Can you please expand PUD and explain what blktap2 does and is used for? can't tell the actors without a playbill, Erik f...@netbsd.org

Perform mmap and poll on PUD character devices

2011-08-31 Thread Roger Pau Monné
Hello, I'm trying to port blktap2 Linux kernel driver to NetBSD using PUD (thanks to libpud developed by Vyacheslav Matyushin during this year SoC). blktap2 is composed of several devices, some of them are block devices, other are character devices. Now I have to implement a character device that