IMHO subject and patch are self explainning :)
--
Andrey Panin | Embedded systems software developer
[email protected] | PGP key: wwwkeys.eu.pgp.net
Andrey Panin wrote:
> diff -urN -X /usr/share/dontdiff linux-vanilla/drivers/net/tulip/de2104x.c linux/drivers/net/tulip/de2104x.c
> --- linux-vanilla/drivers/net/tulip/de2104x.c Sun Sep 1 02:04:53 2002
> +++ linux/drivers/net/tulip/de2104x.c Thu Oct 17 04:10:19 2002
> @@ -2216,7 +2216,7 @@
> .name = DRV_NAME,
> .id_table = de_pci_tbl,
> .probe = de_init_one,
> - .remove = de_remove_one,
> + .remove = __devexit_p(de_remove_one),
> #ifdef CONFIG_PM
> .suspend = de_suspend,
> .resume = de_resume,
alas, it is incorrect, as no one hotplugs this hardware.
Jeff
Jeff Garzik wrote:
>Andrey Panin wrote:
>> diff -urN -X /usr/share/dontdiff linux-vanilla/drivers/net/tulip/de2104x.c \
>> linux/drivers/net/tulip/de2104x.c
>> --- linux-vanilla/drivers/net/tulip/de2104x.c Sun Sep 1 02:04:53 2002
>> +++ linux/drivers/net/tulip/de2104x.c Thu Oct 17 04:10:19 2002
>> @@ -2216,7 +2216,7 @@
>> .name = DRV_NAME,
>> .id_table = de_pci_tbl,
>> .probe = de_init_one,
>> - .remove = de_remove_one,
>> + .remove = __devexit_p(de_remove_one),
>> #ifdef CONFIG_PM
>> .suspend = de_suspend,
>> .resume = de_resume,
>
>
>alas, it is incorrect, as no one hotplugs this hardware.
I believe that there are motherboards that use a chipset from
Compaq that allows hot plugging and unplugging of ordinary PCI cards,
supported by drivers in linux-2.5.43/drivers/hotplug/cpq*.[ch]. At a
trade show, I saw a demo of a motherboard with such a capability (not
running Linux, but I think from Compaq).
So, I believe that all ordinary PCI cards (as opposed to
devices soldered onto motherboards, for example) are now hot plug
capable, although their Linux drivers may not yet be.
Do I misunderstand the situation?
As a side note, I also either do not agree or somehow
misunderstand Jeff Garzik's opposition to devexit_p in non-hotplug
drivers, but that issue will irrelevant in the case of de2104.c if it
is indeed possible for all ordinary PCI form factor cards to be
deployed with the hot plug motherboard chipset that I described.
Adam J. Richter __ ______________ 575 Oroville Road
[email protected] \ / Milpitas, California 95035
+1 408 309-6081 | g g d r a s i l United States of America
"Free Software For The Rest Of Us."
Adam J. Richter wrote:
> I believe that there are motherboards that use a chipset from
> Compaq that allows hot plugging and unplugging of ordinary PCI cards,
> supported by drivers in linux-2.5.43/drivers/hotplug/cpq*.[ch]. At a
> trade show, I saw a demo of a motherboard with such a capability (not
> running Linux, but I think from Compaq).
You are correct that all PCI cards are now hotpluggable.
My position is that _my_ driver will not be converted to be hotpluggable
until someone actually does so. Until such a time, I prefer the space
savings that keeping it non-hotplug-able provides.
Jeff
On Thu, Oct 17, 2002 at 10:21:12AM -0400, Jeff Garzik wrote:
> Andrey Panin wrote:
> >diff -urN -X /usr/share/dontdiff linux-vanilla/drivers/net/tulip/de2104xc
> >linux/drivers/net/tulip/de2104x.c
> >--- linux-vanilla/drivers/net/tulip/de2104x.c Sun Sep 1 02:04:53 2002
> >+++ linux/drivers/net/tulip/de2104x.c Thu Oct 17 04:10:19 2002
> >@@ -2216,7 +2216,7 @@
> > .name = DRV_NAME,
> > .id_table = de_pci_tbl,
> > .probe = de_init_one,
> >- .remove = de_remove_one,
> >+ .remove = __devexit_p(de_remove_one),
> > #ifdef CONFIG_PM
> > .suspend = de_suspend,
> > .resume = de_resume,
>
>
> alas, it is incorrect, as no one hotplugs this hardware.
This patch is not about hotplugging, absence of __devexi_p() makes impossible
to link this driver in kernel.
--
Andrey Panin | Embedded systems software developer
[email protected] | PGP key: wwwkeys.eu.pgp.net
On Thu, 17 Oct 2002, Jeff Garzik wrote:
> Adam J. Richter wrote:
> > I believe that there are motherboards that use a chipset from
> > Compaq that allows hot plugging and unplugging of ordinary PCI cards,
> > supported by drivers in linux-2.5.43/drivers/hotplug/cpq*.[ch]. At a
> > trade show, I saw a demo of a motherboard with such a capability (not
> > running Linux, but I think from Compaq).
>
>
> You are correct that all PCI cards are now hotpluggable.
>
> My position is that _my_ driver will not be converted to be hotpluggable
> until someone actually does so. Until such a time, I prefer the space
> savings that keeping it non-hotplug-able provides.
If I read the original post correctly, you save a LOT of space, becasue
the driver won't link without the patch. Sorry if I misread that, and it's
desirable to avoid such a dependency, but if it's hack or forget, expect
hack.
--
bill davidsen <[email protected]>
CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.