[PATCH] kdump, vmcoreinfo: Export sme_me_mask value to vmcoreinfo

lijiang lijiang at redhat.com
Mon Oct 29 03:12:19 PDT 2018


在 2018年10月29日 17:57, Borislav Petkov 写道:
> On Mon, Oct 29, 2018 at 05:29:16PM +0800, lijiang wrote:
>> Ok, i will change it to a local variable and export it.
> 
> Why do you have to export it at all?!
> 

I mean that i will write the value of local variable to the vmcoreinfo. For example:

diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c
index 4c8acdfdc5a7..3cc975e7ff8f 100644
--- a/arch/x86/kernel/machine_kexec_64.c
+++ b/arch/x86/kernel/machine_kexec_64.c
@@ -352,11 +352,16 @@ void machine_kexec(struct kimage *image)

 void arch_crash_save_vmcoreinfo(void)
 {
+       u64 sme_mask = 0;
+
        VMCOREINFO_NUMBER(phys_base);
        VMCOREINFO_SYMBOL(init_top_pgt);
        vmcoreinfo_append_str("NUMBER(pgtable_l5_enabled)=%d\n",
                        pgtable_l5_enabled());

+       sme_mask = sme_me_mask;
+       VMCOREINFO_NUMBER(sme_mask);
+
 #ifdef CONFIG_NUMA
        VMCOREINFO_SYMBOL(node_data);
        VMCOREINFO_LENGTH(node_data, MAX_NUMNODES);
-- 


Because the crash kernel's page table(pgd/pud/pmd/pte) contains
the memory encryption mask(bit 47), makedumpfile needs to remove
the sme mask to obtain the true physical address.


Thanks.
Lianbo




More information about the kexec mailing list