Re: AtomDis: weird bit sum for PowerPlay?
On Jun 26, 09 18:14:54 +0200, Rafał Miłecki wrote:
Please, check: ./atomdis 95C4.104D.9035.vga.rom d f | grep 3d 00000000: 0006 0410 0c25 003d 003d 0100 .....%.=.=.. 00000020: 003d 003d 0100 0000 0001 0102 0302 0405 .=.=............ 000b: U16 OffsetPowerModeEntries = 0x003d (61) 000d: U16 OffsetPowerUnknownEntries = 0x013d (317) 0021: U16 OffsetPowerModeEntries2 = 0x003d (61) 0023: U16 OffsetPowerUnknownEntries2 = 0x013d (317)
You get some 013d values that are not visible in hex dump. Are you sure this is alright?
All of AtomDis is using the statically defined data structures in
atombios.h (and now atombios_rev.h). Some of the structures (including
this one) isn't static, but parts of the structure sit at positions
depending on offsets that are written earlier in the table.
Dunno how to handle that ATM. We probably need specialized dumpers.
The only way to deal with that is to change the structure in
atombios_rev.h for the device under test. Yes, this sucks.
Matthias
--
Matthias Hopf
I hope to understand types and flags, so I wrote some dumb awk parser for atomdis output to get something readable like: Group 1 (type: 1) modes: [id: 1] engine: 110000 kHz memory: 405000 kHz voltage: 950 flags: 0x0014 [id: 2] engine: 300000 kHz memory: 405000 kHz voltage: 950 flags: 0x0010 [id: 3] engine: 680000 kHz memory: 800000 kHz voltage: 1200 flags: 0x0010 # Usage: # ./atomdis 95C4.104D.9035.vga.rom l | grep PowerPlayInfo # ./atomdis 95C4.104D.9035.vga.rom d f | awk -f ppi2nice.awk Not sure if it's something anyone will use, but as I already have it, posting it :) -- Rafał Miłecki
W dniu 1 lipca 2009 01:01 użytkownik Rafał Miłecki
I hope to understand types and flags, so I wrote some dumb awk parser for atomdis output to get something readable like:
Ups, little mistake in for condition. It didn't display last group. Fixed attached. -- Rafał Miłecki
W dniu 1 lipca 2009 01:19 użytkownik Rafał Miłecki
W dniu 1 lipca 2009 01:01 użytkownik Rafał Miłecki
napisał: I hope to understand types and flags, so I wrote some dumb awk parser for atomdis output to get something readable like:
Ups, little mistake in for condition. It didn't display last group.
Fixed attached.
It didn't work with groups containing other amount of modes than 3. Fixed. I tried to find some relation between modes and flags or groups. No special success here, doesn't make any sense for me. -- Rafał Miłecki
participants (2)
-
Matthias Hopf
-
Rafał Miłecki