[PATCH v3] KEYS: trusted: Use ASN.1 encoded OID

Jarkko Sakkinen jarkko at kernel.org
Thu May 23 14:29:56 UTC 2024


On Thu May 23, 2024 at 5:26 PM EEST, Jarkko Sakkinen wrote:
> On Thu May 23, 2024 at 5:20 PM EEST, Jarkko Sakkinen wrote:
> > There's no reason to encode OID_TPMSealedData at run-time, as it never
> > changes.
> >
> > Replace it with an encoded u8-array, which has the same number of
> > elements:
> >
> > 	67 81 05 0A 01 05
> >
> > Include OBJECT IDENTIFIER (0x06) tag and length as the prologue so that
> > the OID can be simply copied to the blob leading to:
> >
> > 	06 06 67 81 05 0A 01 05
> >
> > Since this in stationary place in the buffer it is guaranteed to always
> > fit and not further checks are required.
> >
> > Reviewed-by: David Howells <dhowells at redhat.com>
> > Signed-off-by: Jarkko Sakkinen <jarkko at kernel.org>
>
> Does not really substitute distribution kernel testing, which is
> IMHO essential for something like TPM2 boot in systemd but for
> simple patches like this, the following does a trivial smoke
> test:
>
> export LINUX_OVERRIDE_SRCDIR=<path to a kernel tree with a trusted keys patch>
> git clone https://gitlab.com/jarkkojs/linux-tpmdd-test.git
> cd linux-tpmdd-test
> cmake -Bbuild && make -Cbuild buildroot-prepare
> make -Cbuild/buildroot/build
> build/buildroot/build/images/run-tests.sh
>
> I'm planning to migrate at some point to systemd and make it
> appear more like distribution tho..
>
> For recompiling just kernel only thing needed is:
>
> rm -rf build/buildroot/build/build/linux-custom
> make -Cbuild/buildroot/build
>
> I've put this also to the MAINTAINERS entry of TPM driver although
> I use it also for keyrings etc. Also it is open for contributions
> via Gitlab merge requests (not requesting them per se but I'm open
> to such possibility).

Right and this is fully CI compatible with both Github and Gitlab
with only a single tweak: BR2_PACKAGE_HOST_QEMU needs to be added
to the qemu config. It is CI agnostic test environment to put 
short.

I've had this also deployed to the CI in the past. It can run both
x86 and aarch64 based runners and even emulates hardware TPM in
three different modes (TPM 1.2, TPM 2.0 FIFO, TPM 2.0 CRB).

I've been just wondering why we need kernel tree changes for Gitlab
CI when you can pretty easily just bootstrap toolchain and qemu
and call it a day (and with trimmed builds it is quite fast too).

BR, Jarkko



More information about the Linux-security-module-archive mailing list