[PATCH] usb hubc build fix.patch prefix

[PATCH] usb hubc build fix.patch prefix

[Paul Jackson]

From: Paul Jackson <pj@sgi.com>

The patch series 2.6.18-mm2-broken-out does not apply to 2.6.18,
for me anyway.

The 'quilt push' of this series fails with:

    Applying patch usb-hubc-build-fix.patch
    patching file drivers/usb/core/hub.c
    Hunk #1 FAILED at 1831.
    Hunk #2 succeeded at 1904 (offset -2 lines).
    Hunk #3 FAILED at 1946.
    2 out of 3 hunks FAILED -- rejects in file drivers/usb/core/hub.c
    Patch usb-hubc-build-fix.patch does not apply (enforce with -f)

If I apply the following patch *just before* the failing
usb-hubc-build-fix.patch, everything applies cleanly from there on
down the patch set.

I don't know what's right here.  I'm just blindly pushing code.

But it seems obvious to me that the 2.6.18-mm2 broken-out patch set
is borked:

  The first patch in the series: origin.patch, definitely places these
  two hub_* defines just before the usb_resume_root_hub() routine.
  
  But then the patch usb-hubc-build-fix.patch clearly expects to find
  those two hub_* defines just before the hub_suspend() routine.

Signed-off-by: Paul Jackson <pj@sgi.com>

---
 drivers/usb/core/hub.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

--- 2.6.18-mm2.orig/drivers/usb/core/hub.c	2006-10-01 17:52:53.000000000 -0700
+++ 2.6.18-mm2/drivers/usb/core/hub.c	2006-10-01 18:26:17.000000000 -0700
@@ -1862,6 +1862,8 @@ static inline int remote_wakeup(struct u
 	return 0;
 }
 
+#define hub_suspend NULL
+#define hub_resume NULL
 #endif
 
 static int hub_suspend(struct usb_interface *intf, pm_message_t msg)
@@ -1946,8 +1948,6 @@ static inline int remote_wakeup(struct u
 	return 0;
 }
 
-#define hub_suspend NULL
-#define hub_resume NULL
 #endif
 
 void usb_resume_root_hub(struct usb_device *hdev)

-- 
                  I won't rest till it's the best ...
                  Programmer, Linux Scalability
                  Paul Jackson <pj@sgi.com> 1.925.600.0401

Re: [PATCH] usb hubc build fix.patch prefix

[Alan Stern]

On Sun, 1 Oct 2006, Paul Jackson wrote:

> From: Paul Jackson <pj@sgi.com>
> 
> The patch series 2.6.18-mm2-broken-out does not apply to 2.6.18,
> for me anyway.
> 
> The 'quilt push' of this series fails with:
> 
>     Applying patch usb-hubc-build-fix.patch
>     patching file drivers/usb/core/hub.c
>     Hunk #1 FAILED at 1831.
>     Hunk #2 succeeded at 1904 (offset -2 lines).
>     Hunk #3 FAILED at 1946.
>     2 out of 3 hunks FAILED -- rejects in file drivers/usb/core/hub.c
>     Patch usb-hubc-build-fix.patch does not apply (enforce with -f)

There is no patch labelled usb-hubc-build-fix.patch anywhere in 
http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18/2.6.18-mm2/broken-out/
This suggests that the quilt archive is indeed messed up.

> If I apply the following patch *just before* the failing
> usb-hubc-build-fix.patch, everything applies cleanly from there on
> down the patch set.
> 
> I don't know what's right here.  I'm just blindly pushing code.
> 
> But it seems obvious to me that the 2.6.18-mm2 broken-out patch set
> is borked:
> 
>   The first patch in the series: origin.patch, definitely places these
>   two hub_* defines just before the usb_resume_root_hub() routine.

That is in fact where they belong.  Furthermore, that's where the 
origin.patch file from the URL above puts them.  It sounds like the 
contents of the archive don't match the contents of the directory.

>   But then the patch usb-hubc-build-fix.patch clearly expects to find
>   those two hub_* defines just before the hub_suspend() routine.

That is wrong.  The module won't compile properly when CONFIG_PM is set 
and CONFIG_USB_SUSPEND isn't if the defines are placed there.

Alan Stern

Re: [PATCH] usb hubc build fix.patch prefix

[Paul Jackson]

Alan wrote:
> There is no patch labelled usb-hubc-build-fix.patch anywhere in 
> http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18/2.6.18-mm2/broken-out/
> This suggests that the quilt archive is indeed messed up.

Thanks - good observation.

I see the same thing here.  The 2.6.18-mm2-broken-out.tar.bz2 I downloaded from

  ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18/2.6.18-mm2/2.6.18-mm2-broken-out.tar.bz2

has a file named broken-out/usb-hubc-build-fix.patch, but the broken out directory

  ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18/2.6.18-mm2/broken-out/

does not have such a file.

-- 
                  I won't rest till it's the best ...
                  Programmer, Linux Scalability
                  Paul Jackson <pj@sgi.com> 1.925.600.0401

Re: [PATCH] usb hubc build fix.patch prefix

[Andrew Morton]

On Mon, 2 Oct 2006 09:27:37 -0700
Paul Jackson <pj@sgi.com> wrote:

> Alan wrote:
> > There is no patch labelled usb-hubc-build-fix.patch anywhere in 
> > http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18/2.6.18-mm2/broken-out/
> > This suggests that the quilt archive is indeed messed up.
> 
> Thanks - good observation.
> 
> I see the same thing here.  The 2.6.18-mm2-broken-out.tar.bz2 I downloaded from
> 
>   ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18/2.6.18-mm2/2.6.18-mm2-broken-out.tar.bz2
> 
> has a file named broken-out/usb-hubc-build-fix.patch, but the broken out directory
> 
>   ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18/2.6.18-mm2/broken-out/
> 
> does not have such a file.
> 

I screwed things up and attempted to fix things by hand post-release and
apparently screwed that up too.  (Matthias mailed me within ten minutes
telling me that the import-mm-into-git script had broken).  

Re: [PATCH] usb hubc build fix.patch prefix

[Paul Jackson]

Andrew wrote:
> I screwed things up and attempted to fix things by hand post-release and ...

Aha.  Looks like we found the culprit ;).

Sounds like you're on it.  Good.

-- 
                  I won't rest till it's the best ...
                  Programmer, Linux Scalability
                  Paul Jackson <pj@sgi.com> 1.925.600.0401