Re: [PATCH v6 05/18] x86,arm,fs/resctrl: Introduce architecture hook to program global kernel mode

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] resctrl: silent truncation of closid in PLZA kmode configuration
--

commit afe4da7391bfbc3b6cf3aee15c764cd25113c4e6
Author: Babu Moger <babu.moger@amd.com>

    x86,arm,fs/resctrl: Introduce architecture hook to program global kernel mode

    RESCTRL_GLOBAL_ENABLE_PER_CPU mode enables kernel work to use different
    allocation and monitoring associations than the originating user task.

diff --git a/arch/x86/kernel/cpu/resctrl/ctrlmondata.c b/arch/x86/kernel/cpu/resctrl/ctrlmondata.c
index e74f1ed54b86c..d5580277db627 100644
--- a/arch/x86/kernel/cpu/resctrl/ctrlmondata.c
+++ b/arch/x86/kernel/cpu/resctrl/ctrlmondata.c
[ ... ]
> +void resctrl_arch_configure_kmode_global(const struct cpumask *cpu_mask,
> +					 u32 closid, bool assign_ctrl,
> +					 u32 rmid, bool assign_mon,
> +					 bool enable)
> +{
> +	union msr_pqr_plza_assoc plza = { 0 };
> +
> +	plza.split.rmid = rmid;
> +	plza.split.rmid_en = assign_mon;
> +	plza.split.closid = closid;

[Severity: Medium]
Will this silently truncate the u32 closid to 4 bits on systems with more
than 16 CLOSIDs?

Looking at union msr_pqr_plza_assoc in
arch/x86/kernel/cpu/resctrl/internal.h, the bitfield restricts closid
to 4 bits:

	struct {
		u64	rmid		:12,
			reserved1	:19,
			rmid_en		: 1,
			closid		: 4,
			reserved2	:11,
			closid_en	: 1,
			reserved3	:15,
			plza_en		: 1;
	} split;

If an administrator assigns a resource group with an index 16 or greater to
kernel mode, the ID will truncate. This causes a QoS resource limit bypass
where kernel execution uses the wrong cache or bandwidth limits.

Given the 11 reserved bits right next to it, is closid: 4 a typo for a
15-bit field?

> +	plza.split.closid_en = assign_ctrl;
> +	plza.split.plza_en = enable;
> +
> +	on_each_cpu_mask(cpu_mask, resctrl_kmode_set_one_amd, &plza, 1);
> +}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/cover.1790867441.git.babu.moger@amd.com?part=5