PATCH: Jeremy's linux 2.6.33-pvops: wrong gntdev

PATCH: Jeremy's linux 2.6.33-pvops: wrong gntdev

[Michael Brade]

[-- Attachment #1.1: Type: Text/Plain, Size: 1033 bytes --]

Hi,

I have compiled Xen and Jeremy's linux-2.6-pvops.git kernel, branch 
"xen/stable-2.6.33.x". This branch uses the wrong gntdev name though, Xen 
says:

xen be core: xen be core: can't open gnttab device

With the following patch Xen works:

diff --git a/drivers/xen/gntdev.c b/drivers/xen/gntdev.c
index ddc59cc..1bd01d4 100644
--- a/drivers/xen/gntdev.c
+++ b/drivers/xen/gntdev.c
@@ -594,7 +594,7 @@ static const struct file_operations gntdev_fops = {
 
 static struct miscdevice gntdev_miscdev = {
        .minor        = MISC_DYNAMIC_MINOR,
-       .name         = "gntdev",
+       .name         = "xen/gntdev",
        .fops         = &gntdev_fops,
 };
 

regards,
  Michael

PS: I am not on this list, so in case you have any comments please CC me.

-- 
Michael Brade;                                              KDE Developer
  |-mail: echo brade !#|tr -d "c oh"|s\e\d 's/e/\@/2;s/$/.org/;s/bra/k/2'
  °--web: http://www.behindkde.org/people/michaelb/

KDE 4: Beyond Your Expectations

[-- Attachment #1.2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

[-- Attachment #2: Type: text/plain, Size: 138 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Re: PATCH: Jeremy's linux 2.6.33-pvops: wrong gntdev

[Pasi Kärkkäinen]

On Sun, Aug 01, 2010 at 09:27:09PM +0200, Michael Brade wrote:
> Hi,
> 
> I have compiled Xen and Jeremy's linux-2.6-pvops.git kernel, branch 
> "xen/stable-2.6.33.x". This branch uses the wrong gntdev name though, Xen 
> says:
> 

I'm not sure if xen/stable-2.6.33.x has all the patches..
xen/stable-2.6.32.x is what most people use..

-- Pasi

> xen be core: xen be core: can't open gnttab device
> 
> With the following patch Xen works:
> 
> diff --git a/drivers/xen/gntdev.c b/drivers/xen/gntdev.c
> index ddc59cc..1bd01d4 100644
> --- a/drivers/xen/gntdev.c
> +++ b/drivers/xen/gntdev.c
> @@ -594,7 +594,7 @@ static const struct file_operations gntdev_fops = {
>  
>  static struct miscdevice gntdev_miscdev = {
>         .minor        = MISC_DYNAMIC_MINOR,
> -       .name         = "gntdev",
> +       .name         = "xen/gntdev",
>         .fops         = &gntdev_fops,
>  };
>  
> 
> regards,
>   Michael
> 
> PS: I am not on this list, so in case you have any comments please CC me.
> 
> -- 
> Michael Brade;                                              KDE Developer
>   |-mail: echo brade !#|tr -d "c oh"|s\e\d 's/e/\@/2;s/$/.org/;s/bra/k/2'
>   °--web: http://www.behindkde.org/people/michaelb/
> 
> KDE 4: Beyond Your Expectations



> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel

Re: PATCH: Jeremy's linux 2.6.33-pvops: wrong gntdev

[Jeremy Fitzhardinge]

  On 08/01/2010 12:27 PM, Michael Brade wrote:
> Hi,
>
> I have compiled Xen and Jeremy's linux-2.6-pvops.git kernel, branch
> "xen/stable-2.6.33.x". This branch uses the wrong gntdev name though, Xen
> says:


Sorry, .33.x is very "stable" in the sense that I haven't really been 
updating it.  As Pasi says, 32.x is where the action is at the moment.  
Once .35 comes out, I plan to make that the new base for development 
(though .32 will be maintained for the long term).

     J

Re: PATCH: Jeremy's linux 2.6.33-pvops: wrong gntdev

[Michael Brade]

[-- Attachment #1.1: Type: Text/Plain, Size: 1010 bytes --]

On Monday 02 August 2010 17:26:30 you wrote:
>   On 08/01/2010 12:27 PM, Michael Brade wrote:
> > Hi,
> > 
> > I have compiled Xen and Jeremy's linux-2.6-pvops.git kernel, branch
> > "xen/stable-2.6.33.x". This branch uses the wrong gntdev name though, Xen
> 
> > says:
> Sorry, .33.x is very "stable" in the sense that I haven't really been
> updating it.  As Pasi says, 32.x is where the action is at the moment.
> Once .35 comes out, I plan to make that the new base for development
> (though .32 will be maintained for the long term).
> 
>      J

Ok, cool, then I'll switch back to .32. Thanks for the info! And hey, you 
know, .35 is out already ;-) I'm really looking forward to that because of the 
XFS improvements.

cheers,
  Michael

-- 
Michael Brade;                                              KDE Developer
  |-mail: echo brade !#|tr -d "c oh"|s\e\d 's/e/\@/2;s/$/.org/;s/bra/k/2'
  °--web: http://www.behindkde.org/people/michaelb/

KDE 4: Beyond Your Expectations

[-- Attachment #1.2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

[-- Attachment #2: Type: text/plain, Size: 138 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Re: PATCH: Jeremy's linux 2.6.33-pvops: wrong gntdev

[Jeremy Fitzhardinge]

  On 08/04/2010 10:30 AM, Michael Brade wrote:
> Ok, cool, then I'll switch back to .32. Thanks for the info! And hey, you
> know, .35 is out already ;-) I'm really looking forward to that because of the
> XFS improvements.

Ah, you use xfs from within a domain?  Does that work OK?

     J

Re: PATCH: Jeremy's linux 2.6.33-pvops: wrong gntdev

[Michael Brade]

[-- Attachment #1.1: Type: Text/Plain, Size: 1807 bytes --]

On Wednesday 04 August 2010 19:54:53 Jeremy Fitzhardinge wrote:
>   On 08/04/2010 10:30 AM, Michael Brade wrote:
> > Ok, cool, then I'll switch back to .32. Thanks for the info! And hey, you
> > know, .35 is out already ;-) I'm really looking forward to that because
> > of the XFS improvements.
> 
> Ah, you use xfs from within a domain?  Does that work OK?

In Dom0, absolutely! I even have a root file system on NFS with XFS using PXE 
in Dom0 and that is running without any problems so far. Just a little slow 
for my taste when many small files need to be processed. Say, for instance, you 
read a tar.gz file from the Dom0-NFS share, unpack it, and write the result 
directly back to that NFS share. That takes a lot longer than it would take 
locally, even with a Gigabit-LAN. But I hope that changes a little with 2.6.35 
and the delayed logging.

The Windows-DomU (hvm-stubdom) uses a file as its harddisk on XFS and also 
seems to work fine (though I am not really using it much, I have it just 
because it's possible :)

The other DomUs I am currently setting up, so I can't say much about that yet. 
There will be a mail server and a web server running on two of those domains, 
both of them using xfs. Should there be any issues I'll let you know.

XFS has been my filesystem of choice for at least 5 years I think. Never had 
any issues with it. Before I used reiserfs and before that I had ext2/3 but 
that one gave me too many problems so I never tried it again.

Would you recommend anything else instead?

cheers,
-- 
Michael Brade;                                              KDE Developer
  |-mail: echo brade !#|tr -d "c oh"|s\e\d 's/e/\@/2;s/$/.org/;s/bra/k/2'
  °--web: http://www.behindkde.org/people/michaelb/

KDE 4: Beyond Your Expectations

[-- Attachment #1.2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

[-- Attachment #2: Type: text/plain, Size: 138 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Re: PATCH: Jeremy's linux 2.6.33-pvops: wrong gntdev

[Jeremy Fitzhardinge]

  On 08/04/2010 11:43 AM, Michael Brade wrote:
> XFS has been my filesystem of choice for at least 5 years I think. Never had
> any issues with it. Before I used reiserfs and before that I had ext2/3 but
> that one gave me too many problems so I never tried it again.
>
> Would you recommend anything else instead?

I ask because xfs does some memory mapping tricks which are actually 
fairly awkward to handle in a PV Xen domain.  We end up disabling its 
lazy vunmap behaviour, which means there's a bit more overhead from tlb 
flushes and the like.  I don't know what kind of performance effect that 
actually has, or in what circumstances.

     J