Weirdness in patch

Weirdness in patch

[Alan Stern]

Greg:

In my copy of gregkh-all-2.6.33.patch (which is dated 2010-03-01), 
there are changes to drivers/base/power/main.c, coming from the 
gregkh-05-driver-core/driver-core-create-lock-unlock-functions-for-struct-device 
patch.

The first weird thing is this.  That file as it currently exists on
www.kernel.org (dated 2010-03-03) contains the following hunk (it's the
last hunk for drivers/base/power/main.c):

@@ -962,7 +962,7 @@ static int device_prepare(struct device
 		suspend_report_result(dev->class->pm->prepare, error);
 	}
  End:
-	up(&dev->sem);
+	device_unlock(dev);
 
 	return error;
 }

But in my copy of gregkh-all-2.6.33.patch, the corresponding hunk is:

@@ -852,7 +852,7 @@
 		suspend_report_result(dev->class->pm->prepare, error);
 	}
  End:
-	up(&dev->sem);
+	device_lock(dev);
 
 	return error;
 }

The device_unlock() got changed to device_lock()!  That messed up my 
suspend testing for a while...

Secondly, none of these changes is present in the 
gregkh-all-2.6.33.patch file at www.kernel.org.  Isn't that supposed to 
contain all the patches from the subsystems merged into one?

What's the story?

Alan Stern

Re: Weirdness in patch

[Greg KH]

On Thu, Mar 04, 2010 at 03:22:22PM -0500, Alan Stern wrote:
> Greg:
> 
> In my copy of gregkh-all-2.6.33.patch (which is dated 2010-03-01), 
> there are changes to drivers/base/power/main.c, coming from the 
> gregkh-05-driver-core/driver-core-create-lock-unlock-functions-for-struct-device 
> patch.
> 
> The first weird thing is this.  That file as it currently exists on
> www.kernel.org (dated 2010-03-03) contains the following hunk (it's the
> last hunk for drivers/base/power/main.c):
> 
> @@ -962,7 +962,7 @@ static int device_prepare(struct device
>  		suspend_report_result(dev->class->pm->prepare, error);
>  	}
>   End:
> -	up(&dev->sem);
> +	device_unlock(dev);
>  
>  	return error;
>  }
> 
> But in my copy of gregkh-all-2.6.33.patch, the corresponding hunk is:
> 
> @@ -852,7 +852,7 @@
>  		suspend_report_result(dev->class->pm->prepare, error);
>  	}
>   End:
> -	up(&dev->sem);
> +	device_lock(dev);
>  
>  	return error;
>  }
> 
> The device_unlock() got changed to device_lock()!  That messed up my 
> suspend testing for a while...
> 
> Secondly, none of these changes is present in the 
> gregkh-all-2.6.33.patch file at www.kernel.org.  Isn't that supposed to 
> contain all the patches from the subsystems merged into one?
> 
> What's the story?

Odd, I don't know.  I changed my script to try to handle git ids instead
of -rc only type base versions, and I probably messed something up.  Let
me go verify that the above patch is still correct (I had to merge it by
hand...)

Yes, the original patch is correct, but somehow the -all patches are not
getting generated correctly.  I'll go look into what broke in my
scripts.  Sorry about that, and thanks for letting me know.

greg k-h

Re: Weirdness in patch

[Greg KH]

On Thu, Mar 04, 2010 at 01:01:19PM -0800, Greg KH wrote:
> Yes, the original patch is correct, but somehow the -all patches are not
> getting generated correctly.  I'll go look into what broke in my
> scripts.  Sorry about that, and thanks for letting me know.

Ok, it all seems to be working now, please let me know if anything looks
wrong after master.kernel.org syncs everything out in about 30 minutes
or so.

thanks,

greg k-h

Re: Weirdness in patch

[Alan Stern]

On Thu, 4 Mar 2010, Greg KH wrote:

> On Thu, Mar 04, 2010 at 01:01:19PM -0800, Greg KH wrote:
> > Yes, the original patch is correct, but somehow the -all patches are not
> > getting generated correctly.  I'll go look into what broke in my
> > scripts.  Sorry about that, and thanks for letting me know.
> 
> Ok, it all seems to be working now, please let me know if anything looks
> wrong after master.kernel.org syncs everything out in about 30 minutes
> or so.

No, it's no good.  The current gregkh-all-2.6.33.patch gets lots of 
errors when I apply it to vanilla 2.6.33.

Alan Stern