2413696_en-US

[#1]

ISSUE: fix dependency for IMX_SEC_ENCLAVE and use-after-free

Hello

This is a follow up from this PR on github.

It seems that the patches provided were at the end not applied as it's not present in lf-6.18.y.

The Kconfig commit is needed to make sure that the Enclave driver cannot be y when NVMEM_IMX_OCOTP_SCU is m, thus avoiding probe defers.

Otherwise you get this:

root@colibri-imx8x-14985125:~# dmesg -l err
[    1.708145] rtc-ds1307 1-0068: hctosys: unable to read the hardware clock
[    2.072996] fsl-se secure-envlave-1: Fail to read FIPS fuse
[    2.078636] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[    2.084513] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[    2.111555] fsl-se secure-envlave-1: Fail to read FIPS fuse
[    2.117209] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[    2.123110] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[    2.141671] fsl-se secure-envlave-1: Fail to read FIPS fuse
[    2.147303] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[    2.153200] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[    2.171418] fsl-se secure-envlave-1: Fail to read FIPS fuse
[    2.177065] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[    2.182929] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[    5.744448] fsl-se secure-envlave-1: Fail to read FIPS fuse
[    5.750217] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[    5.756186] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[    5.877660] fsl-se secure-envlave-1: Fail to read FIPS fuse
[    5.888247] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[    5.894232] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[   10.324138] fsl-se secure-envlave-1: Fail to read FIPS fuse
[   10.349783] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[   10.374692] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[   11.731697] fsl-se secure-envlave-1: Fail to read FIPS fuse
[   11.738134] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[   11.747692] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[   12.284910] fsl-se secure-envlave-1: Fail to read FIPS fuse
[   12.295720] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[   12.307723] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[   12.410541] fsl-se secure-envlave-1: Fail to read FIPS fuse
[   12.426771] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[   12.443446] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[   12.644550] debugfs: '5f1a0000.phy' already exists in 'regmap'
[   12.655906] debugfs: '5f1a0000.phy' already exists in 'regmap'
[   12.670039] fsl-se secure-envlave-1: Fail to read FIPS fuse
[   12.688813] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[   12.696063] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[   12.765923] fsl-se secure-envlave-1: Fail to read FIPS fuse
[   12.775810] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[   12.786887] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[   12.839515] fsl-se secure-envlave-1: Fail to read FIPS fuse
[   12.847285] fsl-se secure-envlave-1: Failed to fetch SoC Info.
[   12.853801] fsl-se secure-envlave-1: failed[-EPROBE_DEFER] to fetch SoC Info
[   12.936721] nvmem imx-scu-ocotp0: cell mac raw len 6 unaligned to nvmem word size 4
[   12.951328] nvmem imx-scu-ocotp0: cell mac raw len 6 unaligned to nvmem word size 4
[   13.708903] genpd_provider mu_a1: failed to power off resource 214 ret -22
[   16.701904] Bluetooth: hci0: unexpected event for opcode 0x0000
root@colibri-imx8x-14985125:~#

On the second commit submitted, there has been some refactoring in the driver and I'm not sure if this has been addressed by NXP with

commit 9703dfecc735 ("LF-15802: drivers: firmware: imx: Fix the SE driver remove")

Could  maybe the SE team confirm?

kinds regards

Franz

Re: ISSUE: fix dependency for IMX_SEC_ENCLAVE and use-after-free

Hello,

I see that this has not been solved yet, will check with internal software team and confirm about this.

Best regards/Saludos,
Aldo.