linux-kernel - two questions: hugetlb, how to set huge_bootmem_page->phys before gather_bootmem

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <57C6933E.2090907@huawei.com>
Date:   Wed, 31 Aug 2016 16:20:14 +0800
From:   Xishi Qiu <qiuxishi@...wei.com>
To:     <grygorii.strashko@...com>, <santosh.shilimkar@...com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        <beckyb@...nel.crashing.org>
CC:     Linux MM <linux-mm@...ck.org>, LKML <linux-kernel@...r.kernel.org>
Subject: two questions: hugetlb, how to set huge_bootmem_page->phys before
 gather_bootmem_prealloc()

If the system is 32 bit, usually we will have a highmem zone.

I find gather_bootmem_prealloc() will free the huge_bootmem_page and
then prep the new huge page in CONFIG_HIGHMEM.

But alloc_bootmem_huge_page() we will use the beginning of the huge page
to store the huge_bootmem_page struct, so how to set huge_bootmem_page->phys?

commit(ee8f248d266ec6966c0ce6b7dec24de43dcc1b58) add phys addr to struct
huge_bootmem_page

Another question, commit(8b89a1169437541a2a9b62c8f7b1a5c0ceb0fbde)
update the interface, and the following code actually fix a bug too, right?

We should use phys instead of virt when calling free_bootmem_late(),
But it has not reported to stable.

-               free_bootmem_late((unsigned long)m,
-                                 sizeof(struct huge_bootmem_page));
+               memblock_free_late(__pa(m),
+                                  sizeof(struct huge_bootmem_page));

Thanks,
Xishi Qiu