2000-12-22 15:34:50

by Matthias Andree

[permalink] [raw]
Subject: FAIL: 2.2.18 + AA-VM-global-7 + serial 5.05

Hi,

I have a vanilla 2.2.18 that I patch Andrea Arcangeli's VM-global-7
patch (for 2.2.18pre25) on top, as well as I?C 2.5.4, the current
--12-09 IDE.2.2.18 patch and ReiserFS 3.5.28. So far, so good. If I now
patch serial 5.05 on top of that, the kernel itself detects devices, but
does nothing if it's to boot /sbin/init. ctrl-alt-del and Magic SysRq
are both functional and can reboot the machine.

Taking out either the VM-global-7 patch or the serial 5.05 update fixes
this.

I suspect that these patches are mutually incompatible.

Could somebody please have a look at this? I will test or provide more
information as requested.

--
Matthias Andree


2000-12-22 15:52:54

by Andrea Arcangeli

[permalink] [raw]
Subject: Re: FAIL: 2.2.18 + AA-VM-global-7 + serial 5.05

On Fri, Dec 22, 2000 at 03:47:57PM +0100, Matthias Andree wrote:
> I suspect that these patches are mutually incompatible.

did you checked that there are no rejects after patching :)

> Could somebody please have a look at this? I will test or provide more
> information as requested.

Where's serial 5.05 so I can have a look?

Andrea

2000-12-22 17:11:40

by Matthias Andree

[permalink] [raw]
Subject: Re: FAIL: 2.2.18 + AA-VM-global-7 + serial 5.05

Andrea Arcangeli schrieb am Freitag, den 22. Dezember 2000:

> On Fri, Dec 22, 2000 at 03:47:57PM +0100, Matthias Andree wrote:
> > I suspect that these patches are mutually incompatible.
>
> did you checked that there are no rejects after patching :)

Yes, I did, there were none.

I had one patch that required a fuzz factor, but it only has a vendor ID
definition in pci.h. It's the patch by Lukasz Trabinski, Subject "Re:
[patch] 2.2.18 PCI_DEVICE_ID_OXSEMI_16PCI954", Date 2000-12-16 19:51
EST, Archive at
http://www.uwsg.indiana.edu/hypermail/linux/kernel/0012.2/0126.html

That patch has been approved by Theodore Y Tso as being correct.

The other patches I applied did not leave any rejects behind.

> > Could somebody please have a look at this? I will test or provide more
> > information as requested.
>
> Where's serial 5.05 so I can have a look?

It's at http://sourceforge.net/projects/serial/ and since these are down
ATM, I'm keeping a copy at
http://www-dt.e-technik.uni-dortmund.de/~ma/kernelpatches/v2.2/v2.2.18/

Procedure:

1. fetch 2.2.18-fix-serial-5.05-pre.patch (has also been posted here, it
just adds a missing #define to pci.h)
2. fetch serial-5.05.tar.gz
3. patch the kernel with that 2.2.18-fix-serial-5.05-pre.patch, it takes
a high fuzz factor (try patch -p1 -F10)
4. unpack serial-5.05
5. sh install-in-kernel
6. patch with VM-global-2.2.18pre25-7
7. make clean dep bzImage modules, install, boot, and see it hang.

Note my kernel also contains the IDE and I?C patches, should that
matter, if it matters, please tell the corresponding maintainer there's
an incompatibility.

For what it's worth, that www-dt server directory also has my config,
it's named config-ma2.

--
Matthias Andree



2000-12-22 17:28:01

by Stuart MacDonald

[permalink] [raw]
Subject: Re: FAIL: 2.2.18 + AA-VM-global-7 + serial 5.05

From: "Matthias Andree" <[email protected]>
> 3. patch the kernel with that 2.2.18-fix-serial-5.05-pre.patch, it takes
> a high fuzz factor (try patch -p1 -F10)
> 4. unpack serial-5.05

I don't have permission to fetch
2.2.18-fix-serial-5.05-pre.patch
at
http://www-dt.e-technik.uni-dortmund.de/~ma/kernelpatches/v2.2/v2.2.18/

What file does step 3 modify? It's likely this patch is being overwritten
(lost) in step 4. Probably not the source of the problem though.

..Stu


2000-12-22 17:44:54

by Matthias Andree

[permalink] [raw]
Subject: Re: FAIL: 2.2.18 + AA-VM-global-7 + serial 5.05

On Fri, 22 Dec 2000, Stuart MacDonald wrote:

> What file does step 3 modify? It's likely this patch is being overwritten
> (lost) in step 4. Probably not the source of the problem though.

No, it's not being overwritten, but it's most likely not the source of
the problem.

Permissions have been fixed, sorry for the inconvenience.

--
Matthias Andree