Peter Zijlstra
2025-Nov-24 14:47 UTC
linux-next: build failure after merge of the tip tree
On Mon, Nov 24, 2025 at 06:00:52AM -0800, Boqun Feng wrote:> On Mon, Nov 24, 2025 at 01:32:49PM +0100, Peter Zijlstra wrote: > > On Mon, Nov 24, 2025 at 11:16:44PM +1100, Stephen Rothwell wrote: > > > Hi all, > > > > > > After merging the tip tree, today's linux-next build (x86_64 allmodconfig) > > > failed like this: > > > > > > ERROR: modpost: "NULL" [drivers/gpu/nova-core/nova_core.ko] undefined! > > > > > > > Rust :/ > > > > So I have: > > > > $ make O=tmp-build LLVM=-22 allmodconfig > > $ make O=tmp-build LLVM=-22 rustavailable > > Rust is available! > > $ grep -e CONFIG_RUST= -e CONFIG_NOVA= tmp-build/.config > > $ > > > > Help ?!? > > > > [CC nova] > > On my system it seems that RANDSTRUCT=y prevents the enabling of RUST, > but I'm sure Danilo and Alexandre have better ideas about how to enable > nova-core.Ah, indeed, RANDSTRUCT seems to be messing things up. I tried disabling that, but then I need to manually enable random crap to get Nova happy and menuconfig isn't very nice. I really don't think rustavailable is doing a good job here. Strictly speaking yes, its available, but figuring out why things don't work is a giant pain in the arse. Now, lets see if frobbing Kconfig to disqualify RANDSTRUCT will let allmodconfig work better.
Peter Zijlstra
2025-Nov-24 14:49 UTC
linux-next: build failure after merge of the tip tree
On Mon, Nov 24, 2025 at 03:47:14PM +0100, Peter Zijlstra wrote:> On Mon, Nov 24, 2025 at 06:00:52AM -0800, Boqun Feng wrote: > > On Mon, Nov 24, 2025 at 01:32:49PM +0100, Peter Zijlstra wrote: > > > On Mon, Nov 24, 2025 at 11:16:44PM +1100, Stephen Rothwell wrote: > > > > Hi all, > > > > > > > > After merging the tip tree, today's linux-next build (x86_64 allmodconfig) > > > > failed like this: > > > > > > > > ERROR: modpost: "NULL" [drivers/gpu/nova-core/nova_core.ko] undefined! > > > > > > > > > > Rust :/ > > > > > > So I have: > > > > > > $ make O=tmp-build LLVM=-22 allmodconfig > > > $ make O=tmp-build LLVM=-22 rustavailable > > > Rust is available! > > > $ grep -e CONFIG_RUST= -e CONFIG_NOVA= tmp-build/.config > > > $ > > > > > > Help ?!? > > > > > > > [CC nova] > > > > On my system it seems that RANDSTRUCT=y prevents the enabling of RUST, > > but I'm sure Danilo and Alexandre have better ideas about how to enable > > nova-core. > > Ah, indeed, RANDSTRUCT seems to be messing things up. I tried disabling > that, but then I need to manually enable random crap to get Nova happy > and menuconfig isn't very nice. > > I really don't think rustavailable is doing a good job here. Strictly > speaking yes, its available, but figuring out why things don't work is a > giant pain in the arse. > > Now, lets see if frobbing Kconfig to disqualify RANDSTRUCT will let > allmodconfig work better.This... ok now lets see wtf the actual problem is. diff --git a/security/Kconfig.hardening b/security/Kconfig.hardening index 86f8768c63d4..4a5890796ac2 100644 --- a/security/Kconfig.hardening +++ b/security/Kconfig.hardening @@ -282,6 +282,7 @@ config CC_HAS_RANDSTRUCT # Randstruct was first added in Clang 15, but it isn't safe to use until # Clang 16 due to https://github.com/llvm/llvm-project/issues/60349 depends on !CC_IS_CLANG || CLANG_VERSION >= 160000 + depends on BROKEN choice prompt "Randomize layout of sensitive kernel structures"