Re: [PATCH v5 4/8] crash: generic crash hotplug support infrastructure

[Eric DeVolder <eric.devolder@oracle.com>]

Baoquan, a comment below.
eric

On 3/24/22 09:37, Eric DeVolder wrote:
> 
> 
> On 3/24/22 09:33, Baoquan He wrote:
>> On 03/24/22 at 08:53am, Eric DeVolder wrote:
>>> Baoquan,
>>> Thanks, I've offered a minor correction below.
>>> eric
>>>
>>> On 3/24/22 08:49, Baoquan He wrote:
>>>> On 03/24/22 at 09:38pm, Baoquan He wrote:
>>>>> On 03/03/22 at 11:27am, Eric DeVolder wrote:
>>>>>> This patch introduces a generic crash hot plug/unplug infrastructure
>>>>>> for CPU and memory changes. Upon CPU and memory changes, a generic
>>>>>> crash_hotplug_handler() obtains the appropriate lock, does some
>>>>>> important house keeping and then dispatches the hot plug/unplug event
>>>>>> to the architecture specific arch_crash_hotplug_handler(), and when
>>>>>> that handler returns, the lock is released.
>>>>>>
>>>>>> This patch modifies crash_core.c to implement a subsys_initcall()
>>>>>> function that installs handlers for hot plug/unplug events. If CPU
>>>>>> hotplug is enabled, then cpuhp_setup_state() is invoked to register a
>>>>>> handler for CPU changes. Similarly, if memory hotplug is enabled, then
>>>>>> register_memory_notifier() is invoked to install a handler for memory
>>>>>> changes. These handlers in turn invoke the common generic handler
>>>>>> crash_hotplug_handler().
>>>>>>
>>>>>> On the CPU side, cpuhp_setup_state_nocalls() is invoked with parameter
>>>>>> CPUHP_AP_ONLINE_DYN. While this works, when a CPU is being unplugged,
>>>>>> the CPU still shows up in foreach_present_cpu() during the regeneration
>>>>>> of the new CPU list, thus the need to explicitly check and exclude the
>>>>>> soon-to-be offlined CPU in crash_prepare_elf64_headers().
>>>>>>
>>>>>> On the memory side, each un/plugged memory block passes through the
>>>>>> handler. For example, if a 1GiB DIMM is hotplugged, that generate 8
>>>>>> memory events, one for each 128MiB memblock.
>>>>>
>>>>> I rewrite the log as below with my understanding. Hope it's simpler to
>>>>> help people get what's going on here. Please consider to take if it's
>>>>> OK to you or adjust based on this. The code looks good to me.
>>>>>
>>>> Made some tuning:
>>>>
>>>> crash: add generic infrastructure for crash hotplug support
>>>>
>>>> Upon CPU and memory changes, a generic crash_hotplug_handler() is added
>>>> to dispatch the hot plug/unplug event to the architecture specific
>>>> arch_crash_hotplug_handler(). During the process, kexec_mutex need be
>>>> held.
>>>>
>>>> To support cpu hotplug, one callback pair are registered to capture
>>>> KEXEC_CRASH_HP_ADD_CPU and KEXEC_CRASH_HP_REMOVE_CPU events via
>>>> cpuhp_setup_state_nocalls().
>>> s/KEXEC_CRASH_HP_ADD}REMOVE_CPU/CPUHP_AP_ONLINE_DYN/ as the KEXEC_CRASH are the
>>> names I've introduced with this patch?
>>
>> Right.
>>
>> While checking it, I notice hp_action which you don't use actually.
>> Can you reconsider that part of design, the hp_action, the a, b
>> parameter passed to handler?
> 
> Sure I can remove. I initially put in there as this was generic infrastructure and not sure if it 
> would benefit others.
> eric
> 

Actually, I will keep the hp_action as the work by Sourabh Jain for PPC uses the hp_action. I'll 
drop the a and b.

Also, shall I post v6, or are you still looking at patches 7 and 8?

Thanks,
eric

>>
>>>
>>>>
>>>> To support memory hotplug, a notifier crash_memhp_nb is registered to
>>>> memory_chain to watch MEM_ONLINE and MEM_OFFLINE events.
>>>>
>>>> These callbacks and notifier will call crash_hotplug_handler() to handle
>>>> captured event when invoked.
>>>>
>>>>>
>>>>>>
>>>>>> Signed-off-by: Eric DeVolder <eric.devolder@oracle.com>
>>>>>> ---
>>>>>>    include/linux/kexec.h |  16 +++++++
>>>>>>    kernel/crash_core.c   | 108 ++++++++++++++++++++++++++++++++++++++++++
>>>>>>    2 files changed, 124 insertions(+)
>>>>>>
>>>>>> diff --git a/include/linux/kexec.h b/include/linux/kexec.h
>>>>>> index d7b59248441b..b11d75a6b2bc 100644
>>>>>> --- a/include/linux/kexec.h
>>>>>> +++ b/include/linux/kexec.h
>>>>>> @@ -300,6 +300,13 @@ struct kimage {
>>>>>>        /* Information for loading purgatory */
>>>>>>        struct purgatory_info purgatory_info;
>>>>>> +
>>>>>> +#ifdef CONFIG_CRASH_HOTPLUG
>>>>>> +    bool hotplug_event;
>>>>>> +    int offlinecpu;
>>>>>> +    bool elf_index_valid;
>>>>>> +    int elf_index;
>>>>>> +#endif
>>>>>>    #endif
>>>>>>    #ifdef CONFIG_IMA_KEXEC
>>>>>> @@ -316,6 +323,15 @@ struct kimage {
>>>>>>        unsigned long elf_load_addr;
>>>>>>    };
>>>>>> +#ifdef CONFIG_CRASH_HOTPLUG
>>>>>> +void arch_crash_hotplug_handler(struct kimage *image,
>>>>>> +    unsigned int hp_action, unsigned long a, unsigned long b);
>>>>>> +#define KEXEC_CRASH_HP_REMOVE_CPU   0
>>>>>> +#define KEXEC_CRASH_HP_ADD_CPU      1
>>>>>> +#define KEXEC_CRASH_HP_REMOVE_MEMORY 2
>>>>>> +#define KEXEC_CRASH_HP_ADD_MEMORY   3
>>>>>> +#endif /* CONFIG_CRASH_HOTPLUG */
>>>>>> +
>>>>>>    /* kexec interface functions */
>>>>>>    extern void machine_kexec(struct kimage *image);
>>>>>>    extern int machine_kexec_prepare(struct kimage *image);
>>>>>> diff --git a/kernel/crash_core.c b/kernel/crash_core.c
>>>>>> index 256cf6db573c..76959d440f71 100644
>>>>>> --- a/kernel/crash_core.c
>>>>>> +++ b/kernel/crash_core.c
>>>>>> @@ -9,12 +9,17 @@
>>>>>>    #include <linux/init.h>
>>>>>>    #include <linux/utsname.h>
>>>>>>    #include <linux/vmalloc.h>
>>>>>> +#include <linux/highmem.h>
>>>>>> +#include <linux/memory.h>
>>>>>> +#include <linux/cpuhotplug.h>
>>>>>>    #include <asm/page.h>
>>>>>>    #include <asm/sections.h>
>>>>>>    #include <crypto/sha1.h>
>>>>>> +#include "kexec_internal.h"
>>>>>> +
>>>>>>    /* vmcoreinfo stuff */
>>>>>>    unsigned char *vmcoreinfo_data;
>>>>>>    size_t vmcoreinfo_size;
>>>>>> @@ -491,3 +496,106 @@ static int __init crash_save_vmcoreinfo_init(void)
>>>>>>    }
>>>>>>    subsys_initcall(crash_save_vmcoreinfo_init);
>>>>>> +
>>>>>> +#ifdef CONFIG_CRASH_HOTPLUG
>>>>>> +void __weak arch_crash_hotplug_handler(struct kimage *image,
>>>>>> +    unsigned int hp_action, unsigned long a, unsigned long b)
>>>>>> +{
>>>>>> +    pr_warn("crash hp: %s not implemented", __func__);
>>>>>> +}
>>>>>> +
>>>>>> +static void crash_hotplug_handler(unsigned int hp_action,
>>>>>> +    unsigned long a, unsigned long b)
>>>>>> +{
>>>>>> +    /* Obtain lock while changing crash information */
>>>>>> +    if (!mutex_trylock(&kexec_mutex))
>>>>>> +        return;
>>>>>> +
>>>>>> +    /* Check kdump is loaded */
>>>>>> +    if (kexec_crash_image) {
>>>>>> +        pr_debug("crash hp: hp_action %u, a %lu, b %lu", hp_action,
>>>>>> +            a, b);
>>>>>> +
>>>>>> +        /* Needed in order for the segments to be updated */
>>>>>> +        arch_kexec_unprotect_crashkres();
>>>>>> +
>>>>>> +        /* Flag to differentiate between normal load and hotplug */
>>>>>> +        kexec_crash_image->hotplug_event = true;
>>>>>> +
>>>>>> +        /* Now invoke arch-specific update handler */
>>>>>> +        arch_crash_hotplug_handler(kexec_crash_image, hp_action, a, b);
>>>>>> +
>>>>>> +        /* No longer handling a hotplug event */
>>>>>> +        kexec_crash_image->hotplug_event = false;
>>>>>> +
>>>>>> +        /* Change back to read-only */
>>>>>> +        arch_kexec_protect_crashkres();
>>>>>> +    }
>>>>>> +
>>>>>> +    /* Release lock now that update complete */
>>>>>> +    mutex_unlock(&kexec_mutex);
>>>>>> +}
>>>>>> +
>>>>>> +#if defined(CONFIG_MEMORY_HOTPLUG)
>>>>>> +static int crash_memhp_notifier(struct notifier_block *nb,
>>>>>> +    unsigned long val, void *v)
>>>>>> +{
>>>>>> +    struct memory_notify *mhp = v;
>>>>>> +    unsigned long start, end;
>>>>>> +
>>>>>> +    start = mhp->start_pfn << PAGE_SHIFT;
>>>>>> +    end = ((mhp->start_pfn + mhp->nr_pages) << PAGE_SHIFT) - 1;
>>>>>> +
>>>>>> +    switch (val) {
>>>>>> +    case MEM_ONLINE:
>>>>>> +        crash_hotplug_handler(KEXEC_CRASH_HP_ADD_MEMORY,
>>>>>> +            start, end-start);
>>>>>> +        break;
>>>>>> +
>>>>>> +    case MEM_OFFLINE:
>>>>>> +        crash_hotplug_handler(KEXEC_CRASH_HP_REMOVE_MEMORY,
>>>>>> +            start, end-start);
>>>>>> +        break;
>>>>>> +    }
>>>>>> +    return NOTIFY_OK;
>>>>>> +}
>>>>>> +
>>>>>> +static struct notifier_block crash_memhp_nb = {
>>>>>> +    .notifier_call = crash_memhp_notifier,
>>>>>> +    .priority = 0
>>>>>> +};
>>>>>> +#endif
>>>>>> +
>>>>>> +#if defined(CONFIG_HOTPLUG_CPU)
>>>>>> +static int crash_cpuhp_online(unsigned int cpu)
>>>>>> +{
>>>>>> +    crash_hotplug_handler(KEXEC_CRASH_HP_ADD_CPU, cpu, 0);
>>>>>> +    return 0;
>>>>>> +}
>>>>>> +
>>>>>> +static int crash_cpuhp_offline(unsigned int cpu)
>>>>>> +{
>>>>>> +    crash_hotplug_handler(KEXEC_CRASH_HP_REMOVE_CPU, cpu, 0);
>>>>>> +    return 0;
>>>>>> +}
>>>>>> +#endif
>>>>>> +
>>>>>> +static int __init crash_hotplug_init(void)
>>>>>> +{
>>>>>> +    int result = 0;
>>>>>> +
>>>>>> +#if defined(CONFIG_MEMORY_HOTPLUG)
>>>>>> +    register_memory_notifier(&crash_memhp_nb);
>>>>>> +#endif
>>>>>> +
>>>>>> +#if defined(CONFIG_HOTPLUG_CPU)
>>>>>> +    result = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN,
>>>>>> +                "crash/cpuhp",
>>>>>> +                crash_cpuhp_online, crash_cpuhp_offline);
>>>>>> +#endif
>>>>>> +
>>>>>> +    return result;
>>>>>> +}
>>>>>> +
>>>>>> +subsys_initcall(crash_hotplug_init);
>>>>>> +#endif /* CONFIG_CRASH_HOTPLUG */
>>>>>> -- 
>>>>>> 2.27.0
>>>>>>
>>>>>
>>>>
>>>
>>