[PATCH v10 8/8] x86/crash: Add x86 crash hotplug support

[Eric DeVolder]

Hi Baoquan,
I've v11 ready to go, but I did raise some questions below which would be good to resolve before 
posting.
Thanks!
eric

On 8/16/22 10:23, Eric DeVolder wrote:
> 
> 
> On 8/12/22 19:34, Baoquan He wrote:
>> On 07/21/22 at 02:17pm, Eric DeVolder wrote:
>> ...snip....
>>> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
>>> index e58798f636d4..bb59596c8bea 100644
>>> --- a/arch/x86/Kconfig
>>> +++ b/arch/x86/Kconfig
>>> @@ -2065,6 +2065,17 @@ config CRASH_DUMP
>>>         (CONFIG_RELOCATABLE=y).
>>>         For more details see Documentation/admin-guide/kdump/kdump.rst
>>> +config CRASH_MAX_MEMORY_RANGES
>>> +    depends on CRASH_DUMP && KEXEC_FILE && (HOTPLUG_CPU || MEMORY_HOTPLUG)
>>> +    int
>>> +    default 32768
>>
>> Do we need to enforce the value with page align and minimal size? I
> 
> Are you asking about the value CRASH_MAX_MEMORY_RANGES? This value represents
> the maximum number of memory ranges, and there Elf64_Phdrs, that we need to
> allow for elfcorehdr memory. So I'm not sure what the concern for alignment
> is. I suppose we could also institute a minimum size for this value, say 1024.
> 
>> checked crash_load_segments() in arch/x86/kernel/crash.c, it does the
>> page size aligning in kexec_add_buffer(). And in
>> load_crashdump_segments() of
>> kexec-tools/kexec/arch/i386/crashdump-x86.c, it creates elfcorehdr at
>> below code, the align is 1024, and in generic add_buffer()
>> implementation, it enforces the memsz page aligned, and changes the
>> passed align as page alignment.
>>
>>
>>     elfcorehdr = add_buffer(info, tmp, bufsz, memsz, align, min_base,
>>                                                          max_addr, -1);
>>
>> Maybe we should at least mention this in the help text to notice people.
> 
> Unfortunately I do not yet understand the concern being raised.
> 
>>
>>> +    help
>>> +      For the kexec_file_load path, specify the maximum number of
>>> +      memory regions, eg. as represented by the 'System RAM' entries
>>> +      in /proc/iomem, that the elfcorehdr buffer/segment can accommodate.
>>> +      This value is combined with NR_CPUS and multiplied by Elf64_Phdr
>>> +      size to determine the final buffer size.
>>> +
>>>   config KEXEC_JUMP
>>>       bool "kexec jump"
>>>       depends on KEXEC && HIBERNATION
>>> diff --git a/arch/x86/include/asm/crash.h b/arch/x86/include/asm/crash.h
>>> index 8b6bd63530dc..96051d8e4b45 100644
>>> --- a/arch/x86/include/asm/crash.h
>>> +++ b/arch/x86/include/asm/crash.h
>>> @@ -9,4 +9,24 @@ int crash_setup_memmap_entries(struct kimage *image,
>>>           struct boot_params *params);
>>>   void crash_smp_send_stop(void);
>>> +void *arch_map_crash_pages(unsigned long paddr, unsigned long size);
>>> +#define arch_map_crash_pages arch_map_crash_pages
>>> +
>>> +void arch_unmap_crash_pages(void **ptr);
>>> +#define arch_unmap_crash_pages arch_unmap_crash_pages
>>> +
>>> +void arch_crash_handle_hotplug_event(struct kimage *image,
>>> +        unsigned int hp_action, unsigned int cpu);
>>> +#define arch_crash_handle_hotplug_event arch_crash_handle_hotplug_event
>>> +
>>> +#ifdef CONFIG_HOTPLUG_CPU
>>> +static inline int crash_hotplug_cpu_support(void) { return 1; }
>>> +#define crash_hotplug_cpu_support crash_hotplug_cpu_support
>>> +#endif
>>> +
>>> +#ifdef CONFIG_MEMORY_HOTPLUG
>>> +static inline int crash_hotplug_memory_support(void) { return 1; }
>>> +#define crash_hotplug_memory_support crash_hotplug_memory_support
>>> +#endif
>>> +
>>>   #endif /* _ASM_X86_CRASH_H */
>>> diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c
>>> index 9ceb93c176a6..55dda4fcde6e 100644
>>> --- a/arch/x86/kernel/crash.c
>>> +++ b/arch/x86/kernel/crash.c
>>> @@ -25,6 +25,7 @@
>>>   #include <linux/slab.h>
>>>   #include <linux/vmalloc.h>
>>>   #include <linux/memblock.h>
>>> +#include <linux/highmem.h>
>>>   #include <asm/processor.h>
>>>   #include <asm/hardirq.h>
>>> @@ -397,7 +398,17 @@ int crash_load_segments(struct kimage *image)
>>>       image->elf_headers = kbuf.buffer;
>>>       image->elf_headers_sz = kbuf.bufsz;
>>> +#if defined(CONFIG_HOTPLUG_CPU) || defined(CONFIG_MEMORY_HOTPLUG)
>>> +    /* Ensure elfcorehdr segment large enough for hotplug changes */
>>> +    kbuf.memsz = (CONFIG_NR_CPUS_DEFAULT + CONFIG_CRASH_MAX_MEMORY_RANGES) * sizeof(Elf64_Phdr);
>>
>> Do we need to break the line to 80 chars?
> 
> Sure, I will do so.
> 
>>
>>> +    /* For marking as usable to crash kernel */
>>> +    image->elf_headers_sz = kbuf.memsz;
>>
>> Do we need this code comment?
> 
> Well, it did take me a while to figure this particular item out in order for all
> this code to work right (else the crash kernel would fail at boot time). So I
> think it best to keep this comment.
> 
>>
>>> +    /* Record the index of the elfcorehdr segment */
>>> +    image->elfcorehdr_index = image->nr_segments;
>>
>> And this place?
> 
> Not necessarily needed, but I've found it useful.
> 
>>
>>> +    image->elfcorehdr_index_valid = true;
>>> +#else
>>>       kbuf.memsz = kbuf.bufsz;
>>> +#endif
>>>       kbuf.buf_align = ELF_CORE_HEADER_ALIGN;
>>>       kbuf.mem = KEXEC_BUF_MEM_UNKNOWN;
>>>       ret = kexec_add_buffer(&kbuf);
>>> @@ -412,3 +423,107 @@ int crash_load_segments(struct kimage *image)
>>>       return ret;
>>>   }
>>>   #endif /* CONFIG_KEXEC_FILE */
>>> +
>>> +#if defined(CONFIG_HOTPLUG_CPU) || defined(CONFIG_MEMORY_HOTPLUG)
>>> +void *arch_map_crash_pages(unsigned long paddr, unsigned long size)
>>> +{
>>> +    /*
>>> +     * NOTE: The addresses and sizes passed to this routine have
>>> +     * already been fully aligned on page boundaries. There is no
>>> +     * need for massaging the address or size.
>>> +     */
>>
>> Can we move the code comment above function interface?
> 
> Yes
> 
>>
>>> +    void *ptr = NULL;
>>> +
>>> +    /* NOTE: requires arch_kexec_[un]protect_crashkres() for write access */
>>
>> Do we need this code comment? On ARCH where proctionion is made, we
>> surely need to the protect/unprotect.
> 
> I will remove this; I've mentioned this in handle_hotplug_event() where these
> protect/unprotect functions are called.
> 
>>
>>> +    if (size > 0) {
>>> +        struct page *page = pfn_to_page(paddr >> PAGE_SHIFT);
>>> +
>>> +        ptr = kmap_local_page(page);
>>> +    }
>>> +
>>> +    return ptr;
>>> +}
>>> +
>>> +void arch_unmap_crash_pages(void **ptr)
>>> +{
>>> +    if (ptr) {
>>> +        if (*ptr)
>>> +            kunmap_local(*ptr);
>>> +        *ptr = NULL;
>>> +    }
>>> +}
>>> +
>>> +/**
>>> + * arch_crash_handle_hotplug_event() - Handle hotplug elfcorehdr changes
>>> + * @image: the active struct kimage
>>> + * @hp_action: the hot un/plug action being handled
>>> + * @cpu: when KEXEC_CRASH_HP_ADD/REMOVE_CPU, the cpu affected
>>> + *
>>> + * To accurately reflect hot un/plug changes, the elfcorehdr (which
>>> + * is passed to the crash kernel via the elfcorehdr= parameter)
>>> + * must be updated with the new list of CPUs and memories. The new
>>> + * elfcorehdr is prepared in a kernel buffer, and then it is
>>> + * written on top of the existing/old elfcorehdr.
>>> + *
>>> + * For hotplug changes to elfcorehdr to work, two conditions are
>>> + * needed:
>>> + * First, the segment containing the elfcorehdr must be large enough
>>> + * to permit a growing number of resources. See the
>>> + * CONFIG_CRASH_MAX_MEMORY_RANGES description.
>>> + * Second, purgatory must explicitly exclude the elfcorehdr from the
>>> + * list of segments it checks (since the elfcorehdr changes and thus
>>> + * would require an update to purgatory itself to update the digest).
>>
>> Isn't this generic concept to crash hotplug? Should we move it out to
>> some generic place?
> 
> Yes, so I will relocate this.
> 
>>
>>> + *
>>> + */
>>> +void arch_crash_handle_hotplug_event(struct kimage *image,
>>> +    unsigned int hp_action, unsigned int cpu)
>>
>> The passed in 'cpu' is not used at all, what is it added for? I didn't
>> see explanation about it.
> 
> Well its not used for x86, but as I recall, Sourabh Jain needed it for the PowerPC handler.
> 
>>
>>> +{
>>> +    struct kexec_segment *ksegment;
>>> +    unsigned char *ptr = NULL;
>>> +    unsigned long elfsz = 0;
>>> +    void *elfbuf = NULL;
>>> +    unsigned long mem, memsz;
>>> +
>>> +    /*
>>> +     * Elfcorehdr_index_valid checked in crash_core:handle_hotplug_event()
>>> +     */
>>> +    ksegment = &image->segment[image->elfcorehdr_index];
>>> +    mem = ksegment->mem;
>>> +    memsz = ksegment->memsz;
>>> +
>>> +    /*
>>> +     * Create the new elfcorehdr reflecting the changes to CPU and/or
>>> +     * memory resources.
>>> +     */
>>> +    if (prepare_elf_headers(image, &elfbuf, &elfsz)) {
>>> +        pr_err("crash hp: unable to prepare elfcore headers");
>>> +        goto out;
>>> +    }
>>> +    if (elfsz > memsz) {
>>> +        pr_err("crash hp: update elfcorehdr elfsz %lu > memsz %lu",
>>> +            elfsz, memsz);
>>> +        goto out;
>>> +    }
>>> +
>>> +    /*
>>> +     * At this point, we are all but assured of success.
>>> +     * Copy new elfcorehdr into destination.
>>> +     */
>>> +    ptr = arch_map_crash_pages(mem, memsz);
>>> +    if (ptr) {
>>> +        /*
>>> +         * Temporarily invalidate the crash image while the
>>> +         * elfcorehdr is updated.
>>> +         */
>>> +        xchg(&kexec_crash_image, NULL);
>>> +        memcpy_flushcache((void *)ptr, elfbuf, elfsz);
>>> +        xchg(&kexec_crash_image, image);
>>> +    }
>>> +    arch_unmap_crash_pages((void **)&ptr);
>>> +    pr_debug("crash hp: re-loaded elfcorehdr at 0x%lx\n", mem);
>>> +
>>> +out:
>>> +    if (elfbuf)
>>> +        vfree(elfbuf);
>>> +}
>>> +#endif
>>> -- 
>>> 2.31.1
>>>
>>