[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <6a2f2abf-4e9a-4569-867c-1aebfb921a0b@amd.com>
Date: Mon, 8 Dec 2025 11:56:34 +1100
From: Alexey Kardashevskiy <aik@....com>
To: Linus Torvalds <torvalds@...ux-foundation.org>, dan.j.williams@...el.com,
Joerg Roedel <joerg.roedel@....com>, Vasant Hegde <vasant.hegde@....com>,
Tom Lendacky <thomas.lendacky@....com>
Cc: linux-coco@...ts.linux.dev, linux-pci@...r.kernel.org,
linux-kernel@...r.kernel.org, Xu Yilun <yilun.xu@...ux.intel.com>,
"Aneesh Kumar K.V (Arm)" <aneesh.kumar@...nel.org>
Subject: Re: [GIT PULL] PCIe Link Encryption and Device Authentication
On 7/12/25 06:21, Linus Torvalds wrote:
> On Fri, 5 Dec 2025 at 19:08, <dan.j.williams@...el.com> wrote:
>>
>> Alexey Kardashevskiy (4):
>> iommu/amd: Report SEV-TIO support
>
> Bah, I've merged this and pushed things out, because my allmodconfig
> build was fine.
>
> But more testing shows that this is broken.
>
> The amd_iommu_sev_tio_supported() function is only defined for
> CONFIG_KVM_AMD_SEV, but the <linux/amd-iommu.h> header put it inside
> the CONFIG_AMD_IOMMU config option block.
>
> So if you have AMD_IOMMU enabled without KVM_AMD_SEV you end up with a
> broken build:
>
> ERROR: modpost: "amd_iommu_sev_tio_supported"
> [drivers/crypto/ccp/ccp.ko] undefined!
> make[2]: *** [scripts/Makefile.modpost:147: Module.symvers] Error 1
>
> I've pushed out a minimal fix that seems to work for me.
>
> Please check - and be more careful. This is _not_ some kind of odd config.
thanks for fixing this and apologies for the late bugs! buuut...
... how do you find these in the first place, just by random configs? It would miss a lot. Dependencies are so freaking complicated (with all these "select" and "depends on" which auto enable/disable options) so trying them all is going to take years :-/ Any advice? Thanks,
>
> Linus
--
Alexey
Powered by blists - more mailing lists