Re: [PATCH v6 3/7] crash_dump: store dm crypt keys in kdump reserved memory

[Coiby Xu]

On Wed, Dec 11, 2024 at 08:58:27PM +0800, Baoquan He wrote:
> On 10/29/24 at 01:52pm, Coiby Xu wrote:
> ......
> > +int crash_load_dm_crypt_keys(struct kimage *image)
> > +{
> > +	struct kexec_buf kbuf = {
> > +		.image = image,
> > +		.buf_min = 0,
> > +		.buf_max = ULONG_MAX,
> > +		.top_down = false,
> > +		.random = true,
> > +	};
> > +	int r;
> > +
> > +
> > +	if (key_count <= 0) {
> > +		kexec_dprintk("No dm-crypt keys\n");
> > +		return -EINVAL;
> > +	}
> > +
> > +	image->dm_crypt_keys_addr = 0;
> > +	r = build_keys_header();
> > +	if (r)
> > +		return r;
> > +
> > +	kbuf.buffer = keys_header;
> > +	kbuf.bufsz = get_keys_header_size(key_count);
> > +
> > +	kbuf.memsz = kbuf.bufsz;
> > +	kbuf.buf_align = ELF_CORE_HEADER_ALIGN;
> > +	kbuf.mem = KEXEC_BUF_MEM_UNKNOWN;
> > +	r = kexec_add_buffer(&kbuf);
> > +	if (r) {
> > +		kvfree((void *)kbuf.buffer);
> > +		return r;
> > +	}
> > +	image->dm_crypt_keys_addr = kbuf.mem;
> > +	image->dm_crypt_keys_sz = kbuf.bufsz;
>
> Wondering why not assigning kbuf.memsz, but bufsz.

Thanks for the question! I think there is no need to use kbuf.memsz
after reading the code on setting image->elf_headers_sz. Currently all
arches except for x86 CONFIG_CRASH_HOTPLUG assigns kbuf.bufsz to
image->elf_headers_sz. For x86 CONFIG_CRASH_HOTPLUG, if I understand it
correctly, it assigns kbuf.memsz to kbuf.memsz because elfcorehdr
changes when there is memory/CPU hot-plugging. For dm crypt keys, there
is no so such concern, so assigning kbuf.bufsz shall be OK.

--
Best regards,
Coiby