Hi Deepak, Russell, Jeff,
The CS89x0 Kconfig entry currently breaks the alignment of all
"EISA, VLB, PCI and on board controllers" that follow it in menuconfig.
This patch fixes it.
This bug was introduced in version 2.6.13-rc1. A first, different fix
attempt was made by Deepak Saxena:
http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=712cb1ebb1653538527500165d8382ca48a7fca1
But it seems it was then overwritten by a subsequent (unsigned?) changeset
from Russell King:
http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=e399822da0f99f8486c33c47e7ae0d32151461e5
Whatever, Deepak's fix wasn't actually fixing the problem for me.
I'm not sure my fix is correct (what exactly are the dependencies
we want for this driver?) but at least it fixes the alignment problem.
It would be nice if we could have this fix in 2.6.15, although it
admittedly isn't critical.
This patch also fixes the module name in the help text, which was
incorrect.
Signed-off-by: Jean Delvare <[email protected]>
Cc: Deepak Saxena <[email protected]>
Cc: Russell King <[email protected]>
Cc: Jeff Garzik <[email protected]>
---
drivers/net/Kconfig | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
--- linux-2.6.15-rc7.orig/drivers/net/Kconfig 2005-12-26 19:59:50.000000000 +0100
+++ linux-2.6.15-rc7/drivers/net/Kconfig 2005-12-27 20:57:05.000000000 +0100
@@ -1374,7 +1374,7 @@
config CS89x0
tristate "CS89x0 support"
- depends on (NET_PCI && (ISA || ARCH_IXDP2X01)) || ARCH_PNX0105
+ depends on NET_PCI && (ISA || ARCH_IXDP2X01 || ARCH_PNX0105)
---help---
Support for CS89x0 chipset based Ethernet cards. If you have a
network (Ethernet) card of this type, say Y and read the
@@ -1384,7 +1384,7 @@
To compile this driver as a module, choose M here and read
<file:Documentation/networking/net-modules.txt>. The module will be
- called cs89x.
+ called cs89x0.
config TC35815
tristate "TOSHIBA TC35815 Ethernet support"
--
Jean Delvare
On Tue, Dec 27, 2005 at 09:06:28PM +0100, Jean Delvare wrote:
> The CS89x0 Kconfig entry currently breaks the alignment of all
> "EISA, VLB, PCI and on board controllers" that follow it in menuconfig.
> This patch fixes it.
>
> This bug was introduced in version 2.6.13-rc1. A first, different fix
> attempt was made by Deepak Saxena:
> http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=712cb1ebb1653538527500165d8382ca48a7fca1
>
> But it seems it was then overwritten by a subsequent (unsigned?) changeset
> from Russell King:
> http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=e399822da0f99f8486c33c47e7ae0d32151461e5
It shouldn't have been overwritten - it's probably the result of a
mis-merge. I would appear that Deepak's change went in via one tree
and mine via my own tree.
As far as that change being unsigned, I don't see why the removal of
a previous addition would require a Sign-off. We never signed-off
for undoing BK changesets so I'm merely following the established
modus operandi.
I have no view on your patch since I never use any of the config
tools apart from oldconfig.
--
Russell King
Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/
maintainer of: 2.6 Serial core
Hi Russell,
> On Tue, Dec 27, 2005 at 09:06:28PM +0100, Jean Delvare wrote:
> > The CS89x0 Kconfig entry currently breaks the alignment of all
> > "EISA, VLB, PCI and on board controllers" that follow it in menuconfig.
> > This patch fixes it.
> >
> > This bug was introduced in version 2.6.13-rc1. A first, different fix
> > attempt was made by Deepak Saxena:
> > http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=712cb1ebb1653538527500165d8382ca48a7fca1
> >
> > But it seems it was then overwritten by a subsequent (unsigned?) changeset
> > from Russell King:
> > http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=e399822da0f99f8486c33c47e7ae0d32151461e5
>
> It shouldn't have been overwritten - it's probably the result of a
> mis-merge. I would appear that Deepak's change went in via one tree
> and mine via my own tree.
>
> As far as that change being unsigned, I don't see why the removal of
> a previous addition would require a Sign-off. We never signed-off
> for undoing BK changesets so I'm merely following the established
> modus operandi.
Given the original meaning of these "Signed-off-by" lines (certifying
that the added code belongs to the sender), it certainly makes sense.
However, we tend to now use these lines for a second meaning, which is
tracing all approvers of the patch from the author to the committer.
For that second meaning, signing reverts may make sense.
> I have no view on your patch since I never use any of the config
> tools apart from oldconfig.
Well, oldconfig is affected by the alignment bug as well. Try deleting
the lines of a few drivers depending on NET_PCI and placed after CS89x0
in .config (e.g. CONFIG_DGRS) and run "make oldconfig", you'll see it.
Aside from the alignment bug, what I really need to know is if the new
dependency expression I came up with it OK:
NET_PCI && (ISA || ARCH_IXDP2X01 || ARCH_PNX0105)
If it is, then I guess that there is no point discussing any further
and the patch can be applied.
Thanks,
--
Jean Delvare