28.01.2014 14:00, Peter Chubb wrote:
[]
> Michael> However, can we please know what is this "our operating
> Michael> system", which source tree to use to run tests, and which
> Michael> "linaro branch" you're talking about?
> 
> Which OS is not important, but if you want to know, it's seL4.

Ah. I thought you're referring to your OS tests, not qemu tests.

Now when I know which test it is, I tried it here - from qemu git.

And here, it is not 0k:

======================
FAIL mmap (output difference)
--- mmap.out    2014-01-28 15:09:38.238713979 +0400
+++ test.log    2014-01-28 15:21:00.147125386 +0400
@@ -58,16 +58,16 @@
 === Running test case: mmap.elf -m 4G ===

 Lower memory: 639k
-Upper memory: 3668984k
+Upper memory: 3144696k

 e820 memory map:
 0x0 - 0x9fc00: type 1 [entry size: 20]
 0x9fc00 - 0xa0000: type 2 [entry size: 20]
 0xf0000 - 0x100000: type 2 [entry size: 20]
-0x100000 - 0xdfffe000: type 1 [entry size: 20]
-0xdfffe000 - 0xe0000000: type 2 [entry size: 20]
+0x100000 - 0xbfffe000: type 1 [entry size: 20]
+0xbfffe000 - 0xc0000000: type 2 [entry size: 20]
 0xfffc0000 - 0x100000000: type 2 [entry size: 20]
-0x100000000 - 0x120000000: type 1 [entry size: 20]
+0x100000000 - 0x140000000: type 1 [entry size: 20]

 mmap start:       0x9000
 mmap end:         0x90a8
@@ -77,16 +77,16 @@
 === Running test case: mmap.elf -m 8G ===

 Lower memory: 639k
-Upper memory: 3668984k
+Upper memory: 3144696k

 e820 memory map:
 0x0 - 0x9fc00: type 1 [entry size: 20]
 0x9fc00 - 0xa0000: type 2 [entry size: 20]
 0xf0000 - 0x100000: type 2 [entry size: 20]
-0x100000 - 0xdfffe000: type 1 [entry size: 20]
-0xdfffe000 - 0xe0000000: type 2 [entry size: 20]
+0x100000 - 0xbfffe000: type 1 [entry size: 20]
+0xbfffe000 - 0xc0000000: type 2 [entry size: 20]
 0xfffc0000 - 0x100000000: type 2 [entry size: 20]
-0x100000000 - 0x220000000: type 1 [entry size: 20]
+0x100000000 - 0x240000000: type 1 [entry size: 20]

 mmap start:       0x9000
 mmap end:         0x90a8
======================

However, on upstream 1.7.0 and 1.6.0 it works fine.

It fails when using seabios, however.  With 1.7.0, and current seabios
from jessie, it shows your zeros:

-Upper memory: 130040k
+Upper memory: 0k
...

It also works fine with current qemu in jessie/sid and with previous
seabios.

Now, I've no idea what's going on, will try to dig further and ask around :)

Thanks,

/mjt


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Reply via email to