On Thu, Sep 19, 2024 at 10:43 AM <devel-request@lists.crash-utility.osci.io> wrote:
Date: Thu, 19 Sep 2024 01:23:05 -0000
From: qiwu.chen@transsion.com
Subject: [Crash-utility] Re: [PATCH v2] kmem: fix the determination
        for slab page
To: devel@lists.crash-utility.osci.io
Message-ID: <20240919012305.29184.37668@lists.crash-utility.osci.io>
Content-Type: text/plain; charset="utf-8"

Hi Lianbo,
Thanks for your review!
>
> Can you try to dump the values of the above two variables from
> the dump_vm_table() and dump_offset_table()? We can display their values by
> help -v and help -o options.
>
Sure, I will post patch v3 to display the two variables.
>
>
> When initializing the page_page_type, the above "if" check seems to be
> redundant.
>
Okay, it should be removed.
>

Thanks.
 

>
> Can this page_type_init() be moved to the page_flags_init()?
>
I don't think so, because the page type is an independent field of struct page which has different usage compared with page flags. What do other reviewers think?

Or can it be placed after the page_flags_init()? Just like this:

diff --git a/memory.c b/memory.c
index 967a9cf58959..9340a8366b98 100644
--- a/memory.c
+++ b/memory.c
@@ -1278,6 +1278,8 @@ vm_init(void)
 
        page_flags_init();
 
+       page_type_init();
+
        rss_page_types_init();
 
        vt->flags |= VM_INIT;

Thanks
Lianbo
 

Thanks
Qiwu