Re: [Crash-utility] What is the current state of the 5-level page tables in Crash
Hi Dave,
At 12/20/2017 02:15 AM, Dave Anderson wrote:
----- Original Message -----
> Dear Dave,
>
> I saw your commit
>
> 307e7f35f510("First phase of future support for x86_64 5-level page
> tables.")
>
> in crash git-tree. It is the preparatory for x86_64 5-level page tables.
> There may be more things to be done to support for 5-level.
Right. There needs to be a manner of determining very early on whether
the kernel page tables are using 5-level. And during crash runtime, there
needs to be a manner of determining whether each user-space task is using
4- or 5-level page tables. Also, it may be possible/preferable to
squeeze the 5-level support into the currently existing x86_64_kvtop() and
x86_64_uvtop() functions instead of using the new xxx_5level() function
placeholders that I added in the preparatory patch.
It is really helpful to me.
>
> I hope to know:
>
> 1. What is the current state of the 5-level page table in Crash?
I have done nothing further beyond than the preparatory steps in that commit.
>
> 2. If there are some reason that we should pause the support work of
> 5-level page tables currently.
No, not really. I originally delayed further work because originally it
was enabled by kernel CONFIG, but as I understand it now, it is enabled by
a combination of a kernel CONFIG and a boot command line option? But
I am sorry I just found the CONFIG_X86_5LEVEL, Could you tell me the
name of boot command line option?
since I have never seen or worked with a kernel with it enabled,
I've
Indeed, we only can use it in QEmu now.
not done anything further.
Yes, I see.
>
> 3. Are you already doing it?
No. Are you offering to complete the work?
Aha, I am willing to, But I am not familiar with crash, I will try me
best to do it.
Thanks,
dou