ACK: [SRU B/F/G LRM] [PATCH 0/1] LP#1919315 -- nvidia.ko assembly failures on end user systems

[Tim Gardner]

Acked-by: Tim Gardner <tim.gardner at canonical.com>

Kinda makes my head hurt 'cause I don't see where the BUILD script is 
used after objdump has been removed. But I guess its been tested.

On 3/16/21 6:26 AM, Andy Whitcroft wrote:
> We are seeing assembly failures of nvidia modules leading to signature
> verification issues.  This is caused indirectly by the CONFIG_MODVERSIONS
> option leading to extra (believed benign) commands in the assembly script.
> These do not affect the base assembly in the linux-restricted-modules,
> but in some cases does affect the identicle assembly on end-users system.
> Is it positied that something external to the assembly package is present
> on affected systems.
> 
> This patch drops the extraneous commands on the end-user systems; they
> are required on the primary to ensure we binaries match.
> 
> This has been sucessfully tested on the prime kernels for groovy, focal,
> and bionic; and on the hwe kernels in focal and bionic.
> 
> Proposing for application to all LRM packages.  Note that we only
> need to expedite fixes for 5.4 and later kernels as the config changes
> were not rolled out to 4.15 (yet).
> 
> -apw
> 
> Branch: lp/1919315-assembly-failures-on-end-user-system
> 
> Andy Whitcroft (1):
>    UBUNTU: [Packaging] quiet (nomially) benign errors in BUILD script
> 
>   debian/scripts/dkms-build--nvidia-N | 3 +++
>   1 file changed, 3 insertions(+)
> 

-- 
-----------
Tim Gardner
Canonical, Inc