linux-toradex.git/arch/parisc/kernel/processor.c, branch v6.19

Revert "parisc: Only list existing CPUs in cpu_possible_mask"

2024-02-06T23:13:21+00:00

This reverts commit 0921244f6f4f0d05698b953fe632a99b38907226.

It broke CPU hotplugging because it modifies the __cpu_possible_mask
after bootup, so that it will be different than nr_cpu_ids, which
then effictively breaks the workqueue setup code and triggers crashes
when shutting down CPUs at runtime.

Guenter was the first who noticed the wrong values in __cpu_possible_mask,
since the cpumask Kunit tests were failig.

Reverting this commit fixes both issues, but sadly brings back this
uncritical runtime warning:
register_cpu_capacity_sysctl: too early to get CPU4 device!

Signed-off-by: Helge Deller 
Reported-by: Guenter Roeck 
Link: https://lkml.org/lkml/2024/2/4/146
Link: https://lore.kernel.org/lkml/Zb0mbHlIud_bqftx@slm.duckdns.org/t/
Cc: stable@vger.kernel.org # 6.0+

parisc: Replace strlcpy() with strscpy()

2023-11-18T17:59:28+00:00

strlcpy() reads the entire source buffer first. This read may exceed
the destination size limit. This is both inefficient and can lead
to linear read overflows if a source string is not NUL-terminated[1].
Additionally, it returns the size of the source string, not the
resulting size of the destination string. In an effort to remove strlcpy()
completely[2], replace strlcpy() here with strscpy().

Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strlcpy [1]
Link: https://github.com/KSPP/linux/issues/89 [2]
Cc: "James E.J. Bottomley" 
Cc: Helge Deller 
Cc: Azeem Shaikh 
Cc: linux-parisc@vger.kernel.org
Signed-off-by: Kees Cook 
Signed-off-by: Helge Deller

parisc: Show default CPU PSW.W setting as reported by PDC

2023-10-30T13:54:41+00:00

The last word shows the default PSW.W setting.

Signed-off-by: Helge Deller

parisc: BTLB: Initialize BTLB tables at CPU startup

2023-09-07T07:12:20+00:00

Initialize the BTLB entries when starting up a CPU.
Note that BTLBs are not available on 64-bit CPUs.

Signed-off-by: Helge Deller

parisc: Fix /proc/cpuinfo output for lscpu

2023-08-22T08:24:46+00:00

The lscpu command is broken since commit cab56b51ec0e ("parisc: Fix
device names in /proc/iomem") added the PA pathname to all PA
devices, includig the CPUs.

lscpu parses /proc/cpuinfo and now believes it found different CPU
types since every CPU is listed with an unique identifier (PA
pathname).

Fix this problem by simply dropping the PA pathname when listing the
CPUs in /proc/cpuinfo. There is no need to show the pathname in this
procfs file.

Fixes: cab56b51ec0e ("parisc: Fix device names in /proc/iomem")
Signed-off-by: Helge Deller 
Cc:  # v4.9+

parisc: processor: Include asm/smp.h for init_per_cpu()

2023-08-10T20:32:24+00:00

Fix sparse warning that init_per_cpu() isn't declared.

Signed-off-by: Helge Deller

parisc: processor: Fix kdoc for init_cpu_profiler()

2023-06-30T15:14:15+00:00

Signed-off-by: Helge Deller

parisc: Show MPE/iX model string at bootup

2022-12-19T15:08:52+00:00

Some (mostly 64-bit machines) machines allow to run MPE/iX and report the MPE
model string via firmware call. Enhance the pdc_model_sysmodel() function to
report that model string.
Note that some 32-bit machines like the B160L wrongly report success for the
firmware call, so include a check to prevent showing wrong info.

Signed-off-by: Helge Deller

parisc: fix a crash with multicore scheduler

2022-06-03T07:54:01+00:00

With the kernel 5.18, the system will hang on boot if it is compiled with
CONFIG_SCHED_MC. The last printed message is "Brought up 1 node, 1 CPU".

The crash happens in sd_init
tl->mask (which is cpu_coregroup_mask) returns an empty mask. This happens
	because cpu_topology[0].core_sibling is empty.
Consequently, sd_span is set to an empty mask
sd_id = cpumask_first(sd_span) sets sd_id == NR_CPUS (because the mask is
	empty)
sd->shared = *per_cpu_ptr(sdd->sds, sd_id); sets sd->shared to NULL
	because sd_id is out of range
atomic_inc(&sd->shared->ref); crashes without printing anything

We can fix it by calling reset_cpu_topology() from init_cpu_topology() -
this will initialize the sibling masks on CPUs, so that they're not empty.

This patch also removes the variable "dualcores_found", it is useless,
because during boot, init_cpu_topology is called before
store_cpu_topology. Thus, set_sched_topology(parisc_mc_topology) is never
called. We don't need to call it at all because default_topology in
kernel/sched/topology.c contains the same items as parisc_mc_topology.

Note that we should not call store_cpu_topology() from init_per_cpu()
because it is called too early in the kernel initialization process and it
results in the message "Failure to register CPU0 device". Before this
patch, store_cpu_topology() would exit immediatelly because
cpuid_topo->core id was uninitialized and it was 0.

Signed-off-by: Mikulas Patocka 
Cc: stable@vger.kernel.org	# v5.18
Signed-off-by: Helge Deller

parisc: Merge model and model name into one line in /proc/cpuinfo

2022-05-08T18:01:11+00:00

The Linux tool "lscpu" shows the double amount of CPUs if we have
"model" and "model name" in two different lines in /proc/cpuinfo.
This change combines the model and the model name into one line.

Signed-off-by: Helge Deller 
Cc: stable@vger.kernel.org