Re: [PATCH v16 7/7] x86/crash: add x86 crash hotplug support

From: Eric DeVolder
Date: Mon Jan 09 2023 - 14:45:36 EST




On 1/9/23 01:36, Baoquan He wrote:
On 01/05/23 at 10:17am, Eric DeVolder wrote:
......
@@ -394,10 +409,37 @@ int crash_load_segments(struct kimage *image)
if (ret)
return ret;
- image->elf_headers = kbuf.buffer;
- image->elf_headers_sz = kbuf.bufsz;
+ image->elf_headers = kbuf.buffer;
+ image->elf_headers_sz = kbuf.bufsz;
+ kbuf.memsz = kbuf.bufsz;
+
+ if (IS_ENABLED(CONFIG_CRASH_HOTPLUG)) {
+ /*
+ * Ensure the elfcorehdr segment large enough for hotplug changes.
+ * Start with VMCOREINFO and kernel_map.
+ */
+ unsigned long pnum = 2;
+
+ if (IS_ENABLED(CONFIG_HOTPLUG_CPU))
+ pnum += CONFIG_NR_CPUS_DEFAULT;
+
+ if (IS_ENABLED(CONFIG_MEMORY_HOTPLUG))
+ pnum += CRASH_MAX_MEMORY_RANGES;

Logic of pnum calculating is a little confusing to me. If I only enable
one of CONFIG_HOTPLUG_CPU and CONFIG_MEMORY_HOTPLUG, is it OK? Say I
only enable CONFIG_HOTPLUG_CPU and CONFIG_SMP, on x86_64,
CONFIG_NR_CPUS_DEFAULT will be 64. pnum will be 64, is it OK. Am I miss
anything?

Ah, your understanding is correct, and the issue you point out I need to fix.
Specifically is only one of HOTPLUG_CPU or MEMORY_HOTPLUG is set, then I need
to allow a default for the other. Using the example you have above, there would
not be room for a single memory region, so I need to allow for a default number
of memory regions.

Good find!
eric


+
+ if (pnum < (unsigned long)PN_XNUM) {
+ kbuf.memsz = pnum * sizeof(Elf64_Phdr);
+ kbuf.memsz += sizeof(Elf64_Ehdr);
+
+ image->elfcorehdr_index = image->nr_segments;
+ image->elfcorehdr_index_valid = true;
+
+ /* Mark as usable to crash kernel, else crash kernel fails on boot */
+ image->elf_headers_sz = kbuf.memsz;
+ } else {
+ pr_err("number of Phdrs %lu exceeds max\n", pnum);
+ }
+ }
- kbuf.memsz = kbuf.bufsz;
kbuf.buf_align = ELF_CORE_HEADER_ALIGN;
kbuf.mem = KEXEC_BUF_MEM_UNKNOWN;
ret = kexec_add_buffer(&kbuf);
@@ -412,3 +454,67 @@ int crash_load_segments(struct kimage *image)
return ret;
}
#endif /* CONFIG_KEXEC_FILE */