linux-kernel - How does the handover from boot allocator to real allocator work

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: <450B1631.40307@goop.org>
Date:	Fri, 15 Sep 2006 14:08:01 -0700
From:	Jeremy Fitzhardinge <jeremy@...p.org>
To:	Ingo Molnar <mingo@...e.hu>, Andrew Morton <akpm@...l.org>
CC:	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: How does the handover from boot allocator to real allocator work

Hi Ingo,

I'm digging around trying to work out how the handover from the 
boot-time allocator to the real memory management works. 

I'm missing something important though:  bootmem.c:free_all_bootmem() 
seems to end up just putting all low memory on the freelists.  How does 
this not put the kernel text+data pages on the freelists?  Also, 
presumably things allocated in the bootmem allocator remain allocated 
for the life of the running kernel?

I'm having a problem in my Xen kernel, in which free_init_pages() ends 
up getting a bad_page() warning because pfn 1024 has the PG_buddy bit 
set on it, which was unexpected (this page ends up being in the middle 
of the initdata section).  Page 1024 gets PF_buddy set by 
free_all_bootmem_core(), and it becomes an order 10 page.

I presume this is broken because free_all_bootmem_core() shouldn't be 
putting the kernel text+data into the freelists, but I don't see how 
this is prevented in the normal i386 case.  I'm guessing it's done by 
something like reserve_bootmem(), but I don't see such a call which 
would reserve the kernel itself.

What am I missing?

Thanks,
    J
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/