stable-3.1: iwlwifi broken with nomsi

stable-3.1: iwlwifi broken with nomsi

[Johan Hovold]

Hi Greg, 

I believe commit 63665158282dd3b5d70 (iwlwifi: allow pci_enable_msi fail)
needs to be added to the 3.1 stable tree as the driver fails to probe
with MSI disabled since commit 084dd79172cb3 (iwlagn: move PCI related
operations from probe and remove to PCI layer) without it:

[    2.564901] iwlagn 0000:0e:00.0: PCI INT A -> GSI 18 (level, low) ->
IRQ 18
[    2.564962] iwlagn 0000:0e:00.0: setting latency timer to 64
[    2.564977] iwlagn 0000:0e:00.0: pci_resource_len = 0x00002000
[    2.567136] iwlagn 0000:0e:00.0: pci_resource_base = f8038000
[    2.569274] iwlagn 0000:0e:00.0: HW Revision ID = 0x0
[    2.571375] iwlagn 0000:0e:00.0: pci_enable_msi failed
[    2.571431] iwlagn 0000:0e:00.0: PCI INT A disabled
[    2.575598] iwlagn: probe of 0000:0e:00.0 failed with error -1

The problem was initially reported here:

	https://lkml.org/lkml/2011/10/26/314

but the fix was never marked for stable.

Thanks,
Johan

Re: stable-3.1: iwlwifi broken with nomsi

[Guy, Wey-Yi]

On Fri, 2011-12-02 at 15:18 -0800, Greg KH wrote:
> On Mon, Nov 28, 2011 at 08:59:39PM +0100, Johan Hovold wrote:
> > Hi Greg, 
> > 
> > I believe commit 63665158282dd3b5d70 (iwlwifi: allow pci_enable_msi fail)
> > needs to be added to the 3.1 stable tree as the driver fails to probe
> > with MSI disabled since commit 084dd79172cb3 (iwlagn: move PCI related
> > operations from probe and remove to PCI layer) without it:
> 
> I do not see commit 63665158282dd3b5d70 in Linus's tree, did you get it
> incorrect?
> 
> confused,
> 
it is commit#8a39ef8ba0fa0410d71db8e981e887fe4fdeca88 in Linville's
wireless tree

Thanks
Wey

Re: stable-3.1: iwlwifi broken with nomsi

[Guy, Wey-Yi]

Hi Greg,

On Fri, 2011-12-02 at 15:32 -0800, Greg KH wrote:
> On Fri, Dec 02, 2011 at 02:28:42PM -0800, Guy, Wey-Yi wrote:
> > On Fri, 2011-12-02 at 15:18 -0800, Greg KH wrote:
> > > On Mon, Nov 28, 2011 at 08:59:39PM +0100, Johan Hovold wrote:
> > > > Hi Greg, 
> > > > 
> > > > I believe commit 63665158282dd3b5d70 (iwlwifi: allow pci_enable_msi fail)
> > > > needs to be added to the 3.1 stable tree as the driver fails to probe
> > > > with MSI disabled since commit 084dd79172cb3 (iwlagn: move PCI related
> > > > operations from probe and remove to PCI layer) without it:
> > > 
> > > I do not see commit 63665158282dd3b5d70 in Linus's tree, did you get it
> > > incorrect?
> > > 
> > > confused,
> > > 
> > it is commit#8a39ef8ba0fa0410d71db8e981e887fe4fdeca88 in Linville's
> > wireless tree
> 
> How nice, and what exactly can I do with that at this point in time?
> 
Can I send the patch to you for stable 3.1?

> 
> Come on, please go read Documentation/stable_kernel_rules.txt again for
> how to do this properly...
> 

Sorry, my mistake and did not realize the offending patch is in 3.1
already.

Wey

Re: stable-3.1: iwlwifi broken with nomsi

[Greg KH]

On Mon, Nov 28, 2011 at 08:59:39PM +0100, Johan Hovold wrote:
> Hi Greg, 
> 
> I believe commit 63665158282dd3b5d70 (iwlwifi: allow pci_enable_msi fail)
> needs to be added to the 3.1 stable tree as the driver fails to probe
> with MSI disabled since commit 084dd79172cb3 (iwlagn: move PCI related
> operations from probe and remove to PCI layer) without it:

I do not see commit 63665158282dd3b5d70 in Linus's tree, did you get it
incorrect?

confused,

greg k-h

Re: stable-3.1: iwlwifi broken with nomsi

[Greg KH]

On Fri, Dec 02, 2011 at 02:28:42PM -0800, Guy, Wey-Yi wrote:
> On Fri, 2011-12-02 at 15:18 -0800, Greg KH wrote:
> > On Mon, Nov 28, 2011 at 08:59:39PM +0100, Johan Hovold wrote:
> > > Hi Greg, 
> > > 
> > > I believe commit 63665158282dd3b5d70 (iwlwifi: allow pci_enable_msi fail)
> > > needs to be added to the 3.1 stable tree as the driver fails to probe
> > > with MSI disabled since commit 084dd79172cb3 (iwlagn: move PCI related
> > > operations from probe and remove to PCI layer) without it:
> > 
> > I do not see commit 63665158282dd3b5d70 in Linus's tree, did you get it
> > incorrect?
> > 
> > confused,
> > 
> it is commit#8a39ef8ba0fa0410d71db8e981e887fe4fdeca88 in Linville's
> wireless tree

How nice, and what exactly can I do with that at this point in time?

{sigh}

Come on, please go read Documentation/stable_kernel_rules.txt again for
how to do this properly...

greg k-h

Re: stable-3.1: iwlwifi broken with nomsi

[Greg KH]

On Fri, Dec 02, 2011 at 02:39:44PM -0800, Guy, Wey-Yi wrote:
> Hi Greg,
> 
> On Fri, 2011-12-02 at 15:32 -0800, Greg KH wrote:
> > On Fri, Dec 02, 2011 at 02:28:42PM -0800, Guy, Wey-Yi wrote:
> > > On Fri, 2011-12-02 at 15:18 -0800, Greg KH wrote:
> > > > On Mon, Nov 28, 2011 at 08:59:39PM +0100, Johan Hovold wrote:
> > > > > Hi Greg, 
> > > > > 
> > > > > I believe commit 63665158282dd3b5d70 (iwlwifi: allow pci_enable_msi fail)
> > > > > needs to be added to the 3.1 stable tree as the driver fails to probe
> > > > > with MSI disabled since commit 084dd79172cb3 (iwlagn: move PCI related
> > > > > operations from probe and remove to PCI layer) without it:
> > > > 
> > > > I do not see commit 63665158282dd3b5d70 in Linus's tree, did you get it
> > > > incorrect?
> > > > 
> > > > confused,
> > > > 
> > > it is commit#8a39ef8ba0fa0410d71db8e981e887fe4fdeca88 in Linville's
> > > wireless tree
> > 
> > How nice, and what exactly can I do with that at this point in time?
> > 
> Can I send the patch to you for stable 3.1?

Again:

> > Come on, please go read Documentation/stable_kernel_rules.txt again for
> > how to do this properly...

Please read that and see _HOW_ to do this, and what the requirements are.

greg k-h

Re: stable-3.1: iwlwifi broken with nomsi

[Johan Hovold]

On Fri, Dec 02, 2011 at 03:18:58PM -0800, Greg KH wrote:
> On Mon, Nov 28, 2011 at 08:59:39PM +0100, Johan Hovold wrote:
> > Hi Greg, 
> > 
> > I believe commit 63665158282dd3b5d70 (iwlwifi: allow pci_enable_msi fail)
> > needs to be added to the 3.1 stable tree as the driver fails to probe
> > with MSI disabled since commit 084dd79172cb3 (iwlagn: move PCI related
> > operations from probe and remove to PCI layer) without it:
> 
> I do not see commit 63665158282dd3b5d70 in Linus's tree, did you get it
> incorrect?

Yes, that should have been 8a39ef8ba0fa0410d71db8e981e887fe4fdeca88 in
Linus' tree. 

Sorry about the confusion,
Johan