drivers/eisa/Makefile | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
I put devlist.h into the wrong Makefile macro ("obj") to get it included
in "targets". Put it into "always" so nothing tries to link against it.
Solves CONFIG_EISA=y i386 build failure:
ld: vmlinux.a: member drivers/eisa/devlist.h in archive is not an object
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Closes: https://lore.kernel.org/all/4a8ba1d0-d2d9-41f8-abf1-d45ec8996d10@infradead.org
Fixes: dd09eb0e2cc4 ("EISA: Increase length of device names")
Signed-off-by: Kees Cook <kees@kernel.org>
---
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Alejandro Colomar <alx@kernel.org>
---
drivers/eisa/Makefile | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/eisa/Makefile b/drivers/eisa/Makefile
index f0d6cf7d1f32..552bd9478340 100644
--- a/drivers/eisa/Makefile
+++ b/drivers/eisa/Makefile
@@ -1,7 +1,8 @@
# SPDX-License-Identifier: GPL-2.0
# Makefile for the Linux device tree
-obj-$(CONFIG_EISA) += devlist.h eisa-bus.o
+always-$(CONFIG_EISA) += devlist.h
+obj-$(CONFIG_EISA) += eisa-bus.o
obj-${CONFIG_EISA_PCI_EISA} += pci_eisa.o
# virtual_root.o should be the last EISA root device to initialize,
--
2.34.1
On 4/22/25 7:37 PM, Kees Cook wrote:
> I put devlist.h into the wrong Makefile macro ("obj") to get it included
> in "targets". Put it into "always" so nothing tries to link against it.
> Solves CONFIG_EISA=y i386 build failure:
>
> ld: vmlinux.a: member drivers/eisa/devlist.h in archive is not an object
>
> Reported-by: Randy Dunlap <rdunlap@infradead.org>
> Closes: https://lore.kernel.org/all/4a8ba1d0-d2d9-41f8-abf1-d45ec8996d10@infradead.org
> Fixes: dd09eb0e2cc4 ("EISA: Increase length of device names")
> Signed-off-by: Kees Cook <kees@kernel.org>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Thanks.
> ---
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Cc: Alejandro Colomar <alx@kernel.org>
> ---
> drivers/eisa/Makefile | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/eisa/Makefile b/drivers/eisa/Makefile
> index f0d6cf7d1f32..552bd9478340 100644
> --- a/drivers/eisa/Makefile
> +++ b/drivers/eisa/Makefile
> @@ -1,7 +1,8 @@
> # SPDX-License-Identifier: GPL-2.0
> # Makefile for the Linux device tree
>
> -obj-$(CONFIG_EISA) += devlist.h eisa-bus.o
> +always-$(CONFIG_EISA) += devlist.h
> +obj-$(CONFIG_EISA) += eisa-bus.o
> obj-${CONFIG_EISA_PCI_EISA} += pci_eisa.o
>
> # virtual_root.o should be the last EISA root device to initialize,
--
~Randy
Hi Kees,
On 4/22/25 8:50 PM, Randy Dunlap wrote:
>
>
> On 4/22/25 7:37 PM, Kees Cook wrote:
>> I put devlist.h into the wrong Makefile macro ("obj") to get it included
>> in "targets". Put it into "always" so nothing tries to link against it.
>> Solves CONFIG_EISA=y i386 build failure:
>>
>> ld: vmlinux.a: member drivers/eisa/devlist.h in archive is not an object
>>
>> Reported-by: Randy Dunlap <rdunlap@infradead.org>
>> Closes: https://lore.kernel.org/all/4a8ba1d0-d2d9-41f8-abf1-d45ec8996d10@infradead.org
>> Fixes: dd09eb0e2cc4 ("EISA: Increase length of device names")
>> Signed-off-by: Kees Cook <kees@kernel.org>
>
> Acked-by: Randy Dunlap <rdunlap@infradead.org>
> Tested-by: Randy Dunlap <rdunlap@infradead.org>
>
> Thanks.
>
Please merge this patch. The build failure is killing lots of i386 builds.
Thanks.
>> ---
>> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>> Cc: Alejandro Colomar <alx@kernel.org>
>> ---
>> drivers/eisa/Makefile | 3 ++-
>> 1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/eisa/Makefile b/drivers/eisa/Makefile
>> index f0d6cf7d1f32..552bd9478340 100644
>> --- a/drivers/eisa/Makefile
>> +++ b/drivers/eisa/Makefile
>> @@ -1,7 +1,8 @@
>> # SPDX-License-Identifier: GPL-2.0
>> # Makefile for the Linux device tree
>>
>> -obj-$(CONFIG_EISA) += devlist.h eisa-bus.o
>> +always-$(CONFIG_EISA) += devlist.h
>> +obj-$(CONFIG_EISA) += eisa-bus.o
>> obj-${CONFIG_EISA_PCI_EISA} += pci_eisa.o
>>
>> # virtual_root.o should be the last EISA root device to initialize,
>
--
~Randy
On Tue, May 06, 2025 at 08:54:29AM -0700, Randy Dunlap wrote:
> Hi Kees,
>
> On 4/22/25 8:50 PM, Randy Dunlap wrote:
> >
> >
> > On 4/22/25 7:37 PM, Kees Cook wrote:
> >> I put devlist.h into the wrong Makefile macro ("obj") to get it included
> >> in "targets". Put it into "always" so nothing tries to link against it.
> >> Solves CONFIG_EISA=y i386 build failure:
> >>
> >> ld: vmlinux.a: member drivers/eisa/devlist.h in archive is not an object
> >>
> >> Reported-by: Randy Dunlap <rdunlap@infradead.org>
> >> Closes: https://lore.kernel.org/all/4a8ba1d0-d2d9-41f8-abf1-d45ec8996d10@infradead.org
> >> Fixes: dd09eb0e2cc4 ("EISA: Increase length of device names")
> >> Signed-off-by: Kees Cook <kees@kernel.org>
> >
> > Acked-by: Randy Dunlap <rdunlap@infradead.org>
> > Tested-by: Randy Dunlap <rdunlap@infradead.org>
> >
> > Thanks.
> >
>
> Please merge this patch. The build failure is killing lots of i386 builds.
> Thanks.
Greg, please take this into your -next tree. The commit it fixes isn't
in Linus's tree, so I don't have any way to land it sanely in -next.
-Kees
--
Kees Cook
On Tue, May 06, 2025 at 08:54:29AM -0700, Randy Dunlap wrote:
> Hi Kees,
>
> On 4/22/25 8:50 PM, Randy Dunlap wrote:
> >
> >
> > On 4/22/25 7:37 PM, Kees Cook wrote:
> >> I put devlist.h into the wrong Makefile macro ("obj") to get it included
> >> in "targets". Put it into "always" so nothing tries to link against it.
> >> Solves CONFIG_EISA=y i386 build failure:
> >>
> >> ld: vmlinux.a: member drivers/eisa/devlist.h in archive is not an object
> >>
> >> Reported-by: Randy Dunlap <rdunlap@infradead.org>
> >> Closes: https://lore.kernel.org/all/4a8ba1d0-d2d9-41f8-abf1-d45ec8996d10@infradead.org
> >> Fixes: dd09eb0e2cc4 ("EISA: Increase length of device names")
> >> Signed-off-by: Kees Cook <kees@kernel.org>
> >
> > Acked-by: Randy Dunlap <rdunlap@infradead.org>
> > Tested-by: Randy Dunlap <rdunlap@infradead.org>
> >
> > Thanks.
> >
>
> Please merge this patch. The build failure is killing lots of i386 builds.
> Thanks.
AFAIK, this needs to go via Greg's tree, but I can see if I can send it
to Linus directly.
-Kees
>
> >> ---
> >> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> >> Cc: Alejandro Colomar <alx@kernel.org>
> >> ---
> >> drivers/eisa/Makefile | 3 ++-
> >> 1 file changed, 2 insertions(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/eisa/Makefile b/drivers/eisa/Makefile
> >> index f0d6cf7d1f32..552bd9478340 100644
> >> --- a/drivers/eisa/Makefile
> >> +++ b/drivers/eisa/Makefile
> >> @@ -1,7 +1,8 @@
> >> # SPDX-License-Identifier: GPL-2.0
> >> # Makefile for the Linux device tree
> >>
> >> -obj-$(CONFIG_EISA) += devlist.h eisa-bus.o
> >> +always-$(CONFIG_EISA) += devlist.h
> >> +obj-$(CONFIG_EISA) += eisa-bus.o
> >> obj-${CONFIG_EISA_PCI_EISA} += pci_eisa.o
> >>
> >> # virtual_root.o should be the last EISA root device to initialize,
> >
>
> --
> ~Randy
>
--
Kees Cook
On 5/6/25 11:51 AM, Kees Cook wrote:
> On Tue, May 06, 2025 at 08:54:29AM -0700, Randy Dunlap wrote:
>> Hi Kees,
>>
>> On 4/22/25 8:50 PM, Randy Dunlap wrote:
>>>
>>>
>>> On 4/22/25 7:37 PM, Kees Cook wrote:
>>>> I put devlist.h into the wrong Makefile macro ("obj") to get it included
>>>> in "targets". Put it into "always" so nothing tries to link against it.
>>>> Solves CONFIG_EISA=y i386 build failure:
>>>>
>>>> ld: vmlinux.a: member drivers/eisa/devlist.h in archive is not an object
>>>>
>>>> Reported-by: Randy Dunlap <rdunlap@infradead.org>
>>>> Closes: https://lore.kernel.org/all/4a8ba1d0-d2d9-41f8-abf1-d45ec8996d10@infradead.org
>>>> Fixes: dd09eb0e2cc4 ("EISA: Increase length of device names")
>>>> Signed-off-by: Kees Cook <kees@kernel.org>
>>>
>>> Acked-by: Randy Dunlap <rdunlap@infradead.org>
>>> Tested-by: Randy Dunlap <rdunlap@infradead.org>
>>>
>>> Thanks.
>>>
>>
>> Please merge this patch. The build failure is killing lots of i386 builds.
>> Thanks.
>
> AFAIK, this needs to go via Greg's tree, but I can see if I can send it
> to Linus directly.
>
> -Kees
Greg told me that it belongs to one of your trees, but Greg did
merge the previous eisa/Makefile patch.
>
>>
>>>> ---
>>>> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>>>> Cc: Alejandro Colomar <alx@kernel.org>
>>>> ---
>>>> drivers/eisa/Makefile | 3 ++-
>>>> 1 file changed, 2 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/drivers/eisa/Makefile b/drivers/eisa/Makefile
>>>> index f0d6cf7d1f32..552bd9478340 100644
>>>> --- a/drivers/eisa/Makefile
>>>> +++ b/drivers/eisa/Makefile
>>>> @@ -1,7 +1,8 @@
>>>> # SPDX-License-Identifier: GPL-2.0
>>>> # Makefile for the Linux device tree
>>>>
>>>> -obj-$(CONFIG_EISA) += devlist.h eisa-bus.o
>>>> +always-$(CONFIG_EISA) += devlist.h
>>>> +obj-$(CONFIG_EISA) += eisa-bus.o
>>>> obj-${CONFIG_EISA_PCI_EISA} += pci_eisa.o
>>>>
>>>> # virtual_root.o should be the last EISA root device to initialize,
>>>
>>
>> --
>> ~Randy
>>
>
--
~Randy
On Tue, May 06, 2025 at 12:17:27PM -0700, Randy Dunlap wrote:
>
>
> On 5/6/25 11:51 AM, Kees Cook wrote:
> > On Tue, May 06, 2025 at 08:54:29AM -0700, Randy Dunlap wrote:
> >> Hi Kees,
> >>
> >> On 4/22/25 8:50 PM, Randy Dunlap wrote:
> >>>
> >>>
> >>> On 4/22/25 7:37 PM, Kees Cook wrote:
> >>>> I put devlist.h into the wrong Makefile macro ("obj") to get it included
> >>>> in "targets". Put it into "always" so nothing tries to link against it.
> >>>> Solves CONFIG_EISA=y i386 build failure:
> >>>>
> >>>> ld: vmlinux.a: member drivers/eisa/devlist.h in archive is not an object
> >>>>
> >>>> Reported-by: Randy Dunlap <rdunlap@infradead.org>
> >>>> Closes: https://lore.kernel.org/all/4a8ba1d0-d2d9-41f8-abf1-d45ec8996d10@infradead.org
> >>>> Fixes: dd09eb0e2cc4 ("EISA: Increase length of device names")
> >>>> Signed-off-by: Kees Cook <kees@kernel.org>
> >>>
> >>> Acked-by: Randy Dunlap <rdunlap@infradead.org>
> >>> Tested-by: Randy Dunlap <rdunlap@infradead.org>
> >>>
> >>> Thanks.
> >>>
> >>
> >> Please merge this patch. The build failure is killing lots of i386 builds.
> >> Thanks.
> >
> > AFAIK, this needs to go via Greg's tree, but I can see if I can send it
> > to Linus directly.
> >
> > -Kees
>
> Greg told me that it belongs to one of your trees, but Greg did
> merge the previous eisa/Makefile patch.
Ah, I thought that was in Kees's tree, sorry about that. I'll take this
then, my fault.
greg k-h
© 2016 - 2026 Red Hat, Inc.