Thread (19 messages) 19 messages, 5 authors, 2009-08-10

Re: [PATCH] Do not inline putprops function

From: Michael Ellerman <hidden>
Date: 2009-08-10 01:51:52
Also in: kexec

On Fri, 2009-08-07 at 20:24 +0530, M. Mohan Kumar wrote:
On Fri, Aug 07, 2009 at 08:05:49PM +0530, M. Mohan Kumar wrote:
quoted
Hi,

After enabling EARLY_DEBUG (and DEBUG in some of the files in
arch/powerpc/kernel directory), without forcing the dtstruct variable to 8
byte alignment: 

# ./kexec -e
Starting new kernel
console [udbg0] enabled
 -> early_setup(), dt_ptr: 0x7723000
 -> early_init_devtree(c000000007723000)
Invalid tag 5 scanning flattened device tree !
search "chosen", depth: 0, uname:
Invalid tag 5 scanning flattened device tree !
dt_root_size_cells = 2
dt_root_addr_cells = 2
Invalid tag 5 scanning flattened device tree !
reserving: 128c000 -> 5ec1f7
reserving: 7734000 -> 8cc000
reserving: 7723000 -> f698
Phys. mem: 0
-> move_device_tree
<- move_device_tree
Scanning CPUs ...
Invalid tag 5 scanning flattened device tree !
 <- early_init_devtree()
Probing machine type ...
  pSeries ...
No suitable machine found !


So device-tree is getting corrupted when dtstruct variable is not aligned to
8 byte variable. This problem is not seen with gcc-3.4. Is it compiler
issue? or bug in the code.
I tried writing the device tree to a binary file with and without dt_len and
compared the files after hexdump:
That sounds like a good idea.
0000 0001 2f00 0000 0000 0003 0000 0004                 0000 0001 2f00 0000 0000 0003 0000 0004
0000 0000 0000 0002 0000 0003 0000 0004                 0000 0000 0000 0002 0000 0003 0000 0004
0000 000f 0000 0002 0000 0003 0000 0004                 0000 000f 0000 0002 0000 0003 0000 0004
0000 001b 3ef1 4800 0000 0003 0000 000d                 0000 001b 3ef1 4800 0000 0003 0000 000d
0000 002b 0000 0000 4942 4d2c 3931 3135               | 0000 002b 4942 4d2c 3931 3135 2d35 3035
2d35 3035 0000 0000 0000 0003 0000 0005               | 0000 0000 0000 0003 0000 0005 0000 0036
So the one on the left (which is ?) - has extra padding after 0x2b,
which is the property data length, before the property value
"IBM,9115-505". There shouldn't be any padding there.

cheers

Attachments

Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help