Re: [PATCH v10 3/8] kexec: define functions to map and unmap segments

[Baoquan He]

On 03/17/25 at 06:04pm, steven chen wrote:
> Currently, the kernel behavior during kexec load is to fetch the IMA
> measurements log from TPM PCRs and store it in a buffer. When a kexec 
> reboot is triggered, this stored log buffer is carried over to the second
> kernel. However, the time gap between kexec load and kexec reboot can be
> very long. During this time window, new events extended into TPM PCRs miss
> the chance to be carried over to the second kernel. This results in a 
> mismatch between TPM PCR quotes and the actual IMA measurements list after
> kexec reboot, leading to remote attestation failure.
> 
> To solve this problem, the new design defers reading TPM PCRs content into
> the kexec buffer to the kexec reboot phase, while still allocating the
> necessary buffer at kexec load time because it is not appropriate to
> allocate memory at the kexec reboot moment.
> 
> The content of memory segments carried over to the new kernel during the
> kexec system call can be changed at the kexec 'execute' stage, but the size
> of the memory segments cannot be changed at the kexec 'execute' stage.
> 
> To copy IMA measurement logs during the kexec operation, IMA allocates 
> memory at the kexec 'load' stage and map the segments to the kimage 
> structure. The mapped address will then be used to copy IMA measurements 
> during the kexec 'execute' stage.
> 
> Currently, the mechanism to map and unmap segments to the kimage structure
> is not available to subsystems outside of kexec.
> 
> Implement kimage_map_segment() to enable IMA to map the measurement log 
> list to the kimage structure during the kexec 'load' stage. This function
> takes a kimage pointer, a memory address, and a size, then gathers the
> source pages within the specified address range, creates an array of page
> pointers, and maps these to a contiguous virtual address range. The
> function returns the start virtual address of this range if successful, 
> or NULL on failure.
> 
> Implement kimage_unmap_segment() for unmapping segments using vunmap().
> 
> From: Tushar Sugandhi <tusharsu@xxxxxxxxxxxxxxxxxxx>
> Signed-off-by: Tushar Sugandhi <tusharsu@xxxxxxxxxxxxxxxxxxx>
> Cc: Eric Biederman <ebiederm@xxxxxxxxxxxx>
> Cc: Baoquan He <bhe@xxxxxxxxxx> 
> Cc: Vivek Goyal <vgoyal@xxxxxxxxxx>
> Cc: Dave Young <dyoung@xxxxxxxxxx>
> Signed-off-by: steven chen <chenste@xxxxxxxxxxxxxxxxxxx>
> ---
>  include/linux/kexec.h |  6 +++++
>  kernel/kexec_core.c   | 54 +++++++++++++++++++++++++++++++++++++++++++
>  2 files changed, 60 insertions(+)

The log is a little verbose, otherwise it looks good to me.

Acked-by: Baoquan He <bhe@xxxxxxxxxx>

> 
> diff --git a/include/linux/kexec.h b/include/linux/kexec.h
> index f0e9f8eda7a3..7d6b12f8b8d0 100644
> --- a/include/linux/kexec.h
> +++ b/include/linux/kexec.h
> @@ -467,13 +467,19 @@ extern bool kexec_file_dbg_print;
>  #define kexec_dprintk(fmt, arg...) \
>          do { if (kexec_file_dbg_print) pr_info(fmt, ##arg); } while (0)
>  
> +extern void *kimage_map_segment(struct kimage *image, unsigned long addr, unsigned long size);
> +extern void kimage_unmap_segment(void *buffer);
>  #else /* !CONFIG_KEXEC_CORE */
>  struct pt_regs;
>  struct task_struct;
> +struct kimage;
>  static inline void __crash_kexec(struct pt_regs *regs) { }
>  static inline void crash_kexec(struct pt_regs *regs) { }
>  static inline int kexec_should_crash(struct task_struct *p) { return 0; }
>  static inline int kexec_crash_loaded(void) { return 0; }
> +static inline void *kimage_map_segment(struct kimage *image, unsigned long addr, unsigned long size)
> +{ return NULL; }
> +static inline void kimage_unmap_segment(void *buffer) { }
>  #define kexec_in_progress false
>  #endif /* CONFIG_KEXEC_CORE */
>  
> diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
> index c0bdc1686154..a5e378e1dc7f 100644
> --- a/kernel/kexec_core.c
> +++ b/kernel/kexec_core.c
> @@ -867,6 +867,60 @@ int kimage_load_segment(struct kimage *image,
>  	return result;
>  }
>  
> +void *kimage_map_segment(struct kimage *image,
> +			 unsigned long addr, unsigned long size)
> +{
> +	unsigned long src_page_addr, dest_page_addr = 0;
> +	unsigned long eaddr = addr + size;
> +	kimage_entry_t *ptr, entry;
> +	struct page **src_pages;
> +	unsigned int npages;
> +	void *vaddr = NULL;
> +	int i;
> +
> +	/*
> +	 * Collect the source pages and map them in a contiguous VA range.
> +	 */
> +	npages = PFN_UP(eaddr) - PFN_DOWN(addr);
> +	src_pages = kmalloc_array(npages, sizeof(*src_pages), GFP_KERNEL);
> +	if (!src_pages) {
> +		pr_err("Could not allocate ima pages array.\n");
> +		return NULL;
> +	}
> +
> +	i = 0;
> +	for_each_kimage_entry(image, ptr, entry) {
> +		if (entry & IND_DESTINATION) {
> +			dest_page_addr = entry & PAGE_MASK;
> +		} else if (entry & IND_SOURCE) {
> +			if (dest_page_addr >= addr && dest_page_addr < eaddr) {
> +				src_page_addr = entry & PAGE_MASK;
> +				src_pages[i++] =
> +					virt_to_page(__va(src_page_addr));
> +				if (i == npages)
> +					break;
> +				dest_page_addr += PAGE_SIZE;
> +			}
> +		}
> +	}
> +
> +	/* Sanity check. */
> +	WARN_ON(i < npages);
> +
> +	vaddr = vmap(src_pages, npages, VM_MAP, PAGE_KERNEL);
> +	kfree(src_pages);
> +
> +	if (!vaddr)
> +		pr_err("Could not map ima buffer.\n");
> +
> +	return vaddr;
> +}
> +
> +void kimage_unmap_segment(void *segment_buffer)
> +{
> +	vunmap(segment_buffer);
> +}
> +
>  struct kexec_load_limit {
>  	/* Mutex protects the limit count. */
>  	struct mutex mutex;
> -- 
> 2.25.1
>