linux-kernel - Re: [PATCH] x86: extend e820 ealy_res support 32bit

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

Message-ID: <483C70F9.2030901@goop.org>
Date:	Tue, 27 May 2008 21:37:13 +0100
From:	Jeremy Fitzhardinge <jeremy@...p.org>
To:	Thomas Gleixner <tglx@...utronix.de>
CC:	Yinghai Lu <yhlu.kernel@...il.com>, Ingo Molnar <mingo@...e.hu>,
	"H. Peter Anvin" <hpa@...or.com>,
	Andrew Morton <akpm@...ux-foundation.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] x86: extend e820 ealy_res support 32bit - fix #2

Thomas Gleixner wrote:
> On Sun, 25 May 2008, Yinghai Lu wrote:
>   
>> remove extra -1 in reseve_early calling
>>     panic if can not find space for new RAMDISK
>>
>>     
>
> Thanks, applied on top of the v2 update
>
> 	tglx
>   
This patch is still required to make Xen boot with the e820 changes:

Subject: xen: boot via i386_start_kernel to get early reservations

Boot Xen via i386_start_kernel so that all the early reservations are
made properly; without these, it will start using the kernel and
pagetables as early heap memory, which is a bit suboptimal.

One tricky part is that reserve_early() will just panic if any of the
early reservations overlap any others.  When a Xen domain is built, it
constructs the initial address space as:

  kernel text+data+bss
  initrd
  inital pagetable

Therefore, when reserving the pagetable (from &_end to
init_pg_tables_end), it covers the whole initrd area.  If it then
tries to reserve the initrd, it will panic because of the overlap.

The simple fix here is to reserve INIT_PG_TABLE first, and then only
reserve the ramdisk if it doesn't overlap with the previous
reservations.  A better/more complex fix might be to make
reserve_early() deal with overlapping reservations.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@...rix.com>
---
 arch/x86/kernel/head32.c |    6 ++++--
 arch/x86/xen/enlighten.c |    2 +-
 include/asm-x86/setup.h  |    1 +
 3 files changed, 6 insertions(+), 3 deletions(-)

===================================================================
--- a/arch/x86/kernel/head32.c
+++ b/arch/x86/kernel/head32.c
@@ -66,6 +66,7 @@ void __init i386_start_kernel(void)
 void __init i386_start_kernel(void)
 {
 	reserve_early(__pa_symbol(&_text), __pa_symbol(&_end), "TEXT DATA BSS");
+	reserve_early(__pa_symbol(&_end), init_pg_tables_end, "INIT_PG_TABLE");
 
 #ifdef CONFIG_BLK_DEV_INITRD
 	/* Reserve INITRD */
@@ -73,10 +74,11 @@ void __init i386_start_kernel(void)
 		u64 ramdisk_image = boot_params.hdr.ramdisk_image;
 		u64 ramdisk_size  = boot_params.hdr.ramdisk_size;
 		u64 ramdisk_end   = ramdisk_image + ramdisk_size;
-		reserve_early(ramdisk_image, ramdisk_end, "RAMDISK");
+		if (ramdisk_end < __pa_symbol(&_text) ||
+		    ramdisk_image >= init_pg_tables_end)
+			reserve_early(ramdisk_image, ramdisk_end, "RAMDISK");
 	}
 #endif
-	reserve_early(__pa_symbol(&_end), init_pg_tables_end, "INIT_PG_TABLE");
 
 	reserve_ebda_region();
 
===================================================================
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -1264,5 +1264,5 @@ asmlinkage void __init xen_start_kernel(
 	}
 
 	/* Start the world */
-	start_kernel();
+	i386_start_kernel();
 }
===================================================================
--- a/include/asm-x86/setup.h
+++ b/include/asm-x86/setup.h
@@ -58,6 +58,7 @@ int __init copy_e820_map(struct e820entr
 int __init copy_e820_map(struct e820entry *biosmap, int nr_map);
 void __init add_memory_region(unsigned long long start,
 			      unsigned long long size, int type);
+void __init i386_start_kernel(void);
 
 extern unsigned long init_pg_tables_end;
 


--
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/