The following 31 patches attempt to bring pci_ids.h back into order.
patch-01-31: INTEL
patch-02-31: S2IO
patch-03-31: RDC
patch-04-31: ADDIDATA
patch-05-31: MELLANOX
patch-06-31: TIGON3
patch-07-31: TITAN
patch-08-31: SEALEVEL
patch-09-31: SIIG
patch-10-31: ITE
patch-11-31: ENSONIQ
patch-12-31: MARVELL
patch-13-31: SBE
patch-14-31: VIA
patch-15-31: NVIDIA
patch-16-31: CMD/SII
patch-17-31: QLOGIC
patch-18-31: YAMAHA
patch-19-31: MYLEX
patch-20-31: PROMISE
patch-21-31: HP
patch-22-31: SI
patch-23-31: MATROX
patch-24-31: TRIDENT
patch-25-31: AMD
patch-26-31: IBM
patch-27-31: DEC/COMPAQ
patch-28-31: NS
patch-29-31: LSI
patch-30-31: COMPAQ
patch-31-31: ESS
Note 1: I have not attempted a build with these changes. I have done
a sort -u on both the original and fully patched files and there
are only two lines that differ due to padding the ID number to
4 digits.
544c544
< #define PCI_DEVICE_ID_COMPAQ_CISSC 0x46
---
> #define PCI_DEVICE_ID_COMPAQ_CISSC 0x0046
720c720
< #define PCI_DEVICE_ID_IBM_ICOM_FOUR_PORT_MODEL 0x252
---
> #define PCI_DEVICE_ID_IBM_ICOM_FOUR_PORT_MODEL 0x0252
Note 2: I have not made any changes to the ATI section as this appears
to be sorted into chip families.
Note 3: I have not moved the following:
PCI_VENDOR_ID_TOSHIBA_2 (it is adjacent to PCI_VENDOR_ID_TOSHIBA)
PCI_DEVICE_ID_SEGA_BBA (it seems to be missing a PCI_VENDOR_ID_SEGA line)
PCI_VENDOR_ID_ADDIDATA_OLD (it is adjacent to PCI_VENDOR_ID_ADDIDATA)
On Wed, 30 Jul 2008 18:50:50 +1000 [email protected] wrote:
> From: [email protected]
> To: unlisted-recipients:;;@imap1.linux-foundation.org (no To-header on input)
Something went wrong with the From: and To: lines.
Jesse wasn't cc'ed?
> Subject: [patch 00/31] pci_ids.h, kernel 2.6.27-rc1, sort DEVICE_IDs within VENDOR_IDs
a) the patch title shouldn't include the kernel version - that
information becomes instantly irrelevant (and soon wrong) as soon as
someone merges the patch.
It is sometimes useful to include this information, but it should
be inside [], so the patch recipient can automatically remove it.
(otoh, the patch _should_ be against latest Linus mainline, so
it's not very useful info).
b) all 31 patches had the same title. This causes numerous
problems. Please think up unique and relevant titles for each
patch.
> Date: Wed, 30 Jul 2008 18:50:50 +1000
> Sender: [email protected]
>
> The following 31 patches attempt to bring pci_ids.h back into order.
This patchset will be utterly ghastly for someone to maintain for two
months. The PCI and linux-next trees will be worst hit.
So we either
a) slam it into mainline immediately or
b) ask you to regenerate it during the 2.6.28 merge window or
c) forget the whole idea.
> a) slam it into mainline immediately or
>
> b) ask you to regenerate it during the 2.6.28 merge window or
>
> c) forget the whole idea.
I vote for (a). At some point it needs doing and the sooner we do it the
better. I've scanned all the diffs and they appear correct. As we've seen
from the MAINTAINERS file once you get errors they spread rapidly.
Acked-by: Alan Cox <[email protected]>
On Wed, Jul 30, 2008 at 06:50:50PM +1000, [email protected] wrote:
> The following 31 patches attempt to bring pci_ids.h back into order.
Why not just fold this into one patch? I can't think of a good reason
to do this one vendor at a time. (Aside from someone trying to do
Signed-off-by: Statistics whoring, of course. :-)
- Ted
Sorry about omitting Jesse (and stuffing up the mail headers).
I am happy to regenerate it as a single patch if preferred and resend
fixing all the mistakes I made in the method of submission.
On Wednesday, July 30, 2008 3:35 am Alan Cox wrote:
> > a) slam it into mainline immediately or
> >
> > b) ask you to regenerate it during the 2.6.28 merge window or
> >
> > c) forget the whole idea.
>
> I vote for (a). At some point it needs doing and the sooner we do it the
> better. I've scanned all the diffs and they appear correct. As we've seen
> from the MAINTAINERS file once you get errors they spread rapidly.
>
> Acked-by: Alan Cox <[email protected]>
Yeah, let's just get it over with.
Acked-by: Jesse Barnes <[email protected]>
Thanks,
Jesse
On Wed, 30 Jul 2008 07:23:14 -0700
Jesse Barnes <[email protected]> wrote:
> On Wednesday, July 30, 2008 3:35 am Alan Cox wrote:
> > > a) slam it into mainline immediately or
> > >
> > > b) ask you to regenerate it during the 2.6.28 merge window or
> > >
> > > c) forget the whole idea.
> >
> > I vote for (a). At some point it needs doing and the sooner we do
> > it the better. I've scanned all the diffs and they appear correct.
> > As we've seen from the MAINTAINERS file once you get errors they
> > spread rapidly.
> >
> > Acked-by: Alan Cox <[email protected]>
>
> Yeah, let's just get it over with.
>
if this was done by a script.. can we just get that script so that you
as PCI maintainer can run that on a very regular basis?
On Wednesday, July 30, 2008 8:10 am Arjan van de Ven wrote:
> On Wed, 30 Jul 2008 07:23:14 -0700
>
> Jesse Barnes <[email protected]> wrote:
> > On Wednesday, July 30, 2008 3:35 am Alan Cox wrote:
> > > > a) slam it into mainline immediately or
> > > >
> > > > b) ask you to regenerate it during the 2.6.28 merge window or
> > > >
> > > > c) forget the whole idea.
> > >
> > > I vote for (a). At some point it needs doing and the sooner we do
> > > it the better. I've scanned all the diffs and they appear correct.
> > > As we've seen from the MAINTAINERS file once you get errors they
> > > spread rapidly.
> > >
> > > Acked-by: Alan Cox <[email protected]>
> >
> > Yeah, let's just get it over with.
>
> if this was done by a script.. can we just get that script so that you
> as PCI maintainer can run that on a very regular basis?
Sure, though I'll also try to be vigilant about future additions.
Jesse