FAILED: patch "[PATCH] x86/cacheinfo: Add a cpu_llc_shared_mask() UP variant" failed to apply to 5.15-stable tree

The patch below does not apply to the 5.15-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

Possible dependencies:

df5b035b5683 ("x86/cacheinfo: Add a cpu_llc_shared_mask() UP variant") 66558b730f25 ("sched: Add cluster scheduler level for x86")

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From df5b035b5683d6a25f077af889fb88e09827f8bc Mon Sep 17 00:00:00 2001
From: Borislav Petkov <[email protected]>
Date: Fri, 19 Aug 2022 19:47:44 +0200
Subject: [PATCH] x86/cacheinfo: Add a cpu_llc_shared_mask() UP variant

On a CONFIG_SMP=n kernel, the LLC shared mask is 0, which prevents __cache_amd_cpumap_setup() from doing the L3 masks setup, and more specifically from setting up the shared_cpu_map and shared_cpu_list files in sysfs, leading to lscpu from util-linux getting confused and segfaulting.

Add a cpu_llc_shared_mask() UP variant which returns a mask with a single bit set, i.e., for CPU0.

index 81a0211a372d..a73bced40e24 100644
@@ -21,16 +21,6 @@ DECLARE_PER_CPU_READ_MOSTLY(u16, cpu_llc_id);
 DECLARE_PER_CPU_READ_MOSTLY(u16, cpu_l2c_id);
 DECLARE_PER_CPU_READ_MOSTLY(int, cpu_number);
 
-static inline struct cpumask *cpu_llc_shared_mask(int cpu)
-{
-	return per_cpu(cpu_llc_shared_map, cpu);
-}
-
-static inline struct cpumask *cpu_l2c_shared_mask(int cpu)
-{
-	return per_cpu(cpu_l2c_shared_map, cpu);
-}
-
 DECLARE_EARLY_PER_CPU_READ_MOSTLY(u16, x86_cpu_to_apicid);
 DECLARE_EARLY_PER_CPU_READ_MOSTLY(u32, x86_cpu_to_acpiid);
 DECLARE_EARLY_PER_CPU_READ_MOSTLY(u16, x86_bios_cpu_apicid);

@@ -172,6 +162,16 @@ extern int safe_smp_processor_id(void);
 # define safe_smp_processor_id()	smp_processor_id()
 #endif
 
+static inline struct cpumask *cpu_llc_shared_mask(int cpu)
+{
+	return per_cpu(cpu_llc_shared_map, cpu);
+}
+
+static inline struct cpumask *cpu_l2c_shared_mask(int cpu)
+{
+	return per_cpu(cpu_l2c_shared_map, cpu);
+}
+
 #else /* !CONFIG_SMP */
 #define wbinvd_on_cpu(cpu)     wbinvd()
 static inline int wbinvd_on_all_cpus(void)

@@ -179,6 +179,11 @@ static inline int wbinvd_on_all_cpus(void)
 	wbinvd();
 	return 0;
 }
+
+static inline struct cpumask *cpu_llc_shared_mask(int cpu)
+{
+	return (struct cpumask *)cpumask_of(0);
+}
 #endif /* CONFIG_SMP */
 
 extern unsigned disabled_cpus;

This is a fix for CONFIG_SMP=n kernels which was caught by testing this explicitly by disabling SMP. IOW, I don't think anyone would be running SMP=n kernels and thus maybe should not backport those...?

Thx.

Regards/Gruss,
    Boris.

SUSE Software Solutions Germany GmbH
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman
(HRB 36809, AG Nürnberg)

Yeah, I agree, this doesn't look like a valid thing to backport, if people had this issue in 5.15.y they would have already reported by now.

thanks,

greg k-h