Bug#943953: linux: DKMS module builds are failing on arm64 due to lack of armhf cross-compiler

2019-11-04 Thread Ben Hutchings
Control: tag -1 pending

On Mon, 2019-11-04 at 11:48 +0100, Raphael Hertzog wrote:
> Control: tags -1 - moreinfo
> 
> Hi,
> 
> On Sat, 02 Nov 2019, Ben Hutchings wrote:
> > I already communicated this via IRC, but I thought it should be logged
> > in the BTS too: I have a tentative fix for this in <
> > https://salsa.debian.org/benh/linux.git> branch bug943953.  I don't
> > have an arm64 system to test it on though.
> 
> Steev tested it and replied to you on IRC:
> 03:09  bwh: buxy: tested on two different arm64 machines, seems to be 
> working

So I've merged this change to the sid branch.

Ben.

-- 
Ben Hutchings
If God had intended Man to program,
we'd have been born with serial I/O ports.




signature.asc
Description: This is a digitally signed message part


Bug#943953: linux: DKMS module builds are failing on arm64 due to lack of armhf cross-compiler

2019-11-04 Thread Raphael Hertzog
Control: tags -1 - moreinfo

Hi,

On Sat, 02 Nov 2019, Ben Hutchings wrote:
> I already communicated this via IRC, but I thought it should be logged
> in the BTS too: I have a tentative fix for this in <
> https://salsa.debian.org/benh/linux.git> branch bug943953.  I don't
> have an arm64 system to test it on though.

Steev tested it and replied to you on IRC:
03:09  bwh: buxy: tested on two different arm64 machines, seems to be 
working

Cheers,
-- 
Raphaël Hertzog ◈ Debian Developer

Support Debian LTS: https://www.freexian.com/services/debian-lts.html
Learn to master Debian: https://debian-handbook.info/get/


signature.asc
Description: PGP signature


Bug#943953: linux: DKMS module builds are failing on arm64 due to lack of armhf cross-compiler

2019-11-02 Thread Ben Hutchings
Control: tag -1 moreinfo patch

I already communicated this via IRC, but I thought it should be logged
in the BTS too: I have a tentative fix for this in <
https://salsa.debian.org/benh/linux.git> branch bug943953.  I don't
have an arm64 system to test it on though.

Ben.

-- 
Ben Hutchings
Humans are not rational beings; they are rationalising beings.



signature.asc
Description: This is a digitally signed message part


Bug#943953: linux: DKMS module builds are failing on arm64 due to lack of armhf cross-compiler

2019-11-01 Thread Raphaël Hertzog
Source: linux
Version: 5.3.7-1
Severity: normal
User: de...@kali.org
Usertags: origin-kali

https://salsa.debian.org/kernel-team/linux/commit/82c843e1577930c4eb168f626ab9bd483b118efc
seems to break DKMS on arm64:


root@pinebook-kali:/usr/src# dkms autoinstall

Kernel preparation unnecessary for this kernel.  Skipping...

Building module:
cleaning build area...
'make' -j4 ARCH=arm64 KVER=5.3.0-kali1-arm64 
KSRC=/lib/modules/5.3.0-kali1-arm64/build/...(bad exit status: 2)
Error! Bad return status for module build on kernel: 5.3.0-kali1-arm64 (aarch64)
Consult /var/lib/dkms/rtl8723cs/2019.07.31/build/make.log for more information.
root@pinebook-kali:/usr/src# cat 
/var/lib/dkms/rtl8723cs/2019.07.31/build/make.log 
DKMS make.log for rtl8723cs-2019.07.31 for kernel 5.3.0-kali1-arm64 (aarch64)
Fri Nov  1 03:50:50 CDT 2019
make ARCH=arm64 CROSS_COMPILE= -C /lib/modules/5.3.0-kali1-arm64/build/ 
M=/var/lib/dkms/rtl8723cs/2019.07.31/build  modules
make[1]: Entering directory '/usr/src/linux-headers-5.3.0-kali1-arm64'
arch/arm64/Makefile:58: *** arm-linux-gnueabihf-gcc not found, check 
CROSS_COMPILE_COMPAT.  Stop.
make[1]: *** [/usr/src/linux-headers-5.3.0-kali1-common/Makefile:179: sub-make] 
Error 2
make[1]: Leaving directory '/usr/src/linux-headers-5.3.0-kali1-arm64'
make: *** [Makefile:1670: modules] Error 2


Some more information gathered on IRC:

16:51  Oh, we have the Depends right, so I think the check in 
arch/arm64/Makefile needs to be suppressed when building out-of-tree modules
16:54  bwh: what do you mean with "we have the Depends right" (i.e. with 
or without the cross-compiler)? on what package are you looking?
16:55  The Depends for the linux-headers package don't include an armhf 
compiler, which is correct
16:55  but the kernel build system still looks for it, which is not
16:55  There are some upstream changes around this that might fix it

Cheers,
 Raphaël.