Dear Timothy, dear Damien,
looking through the board status data, I noticed the boot time increased by one second for the ASUS KGPE-D16.
``` asus/kgpe-d16/4.3-584-gcc9963e/2016-03-25T19_52_04Z/coreboot_timestamps.txt:Total Time: 18,389,685 asus/kgpe-d16/4.3-587-gec38c3d/2016-03-26T21_40_56Z/coreboot_timestamps.txt:Total Time: 19,425,140 ```
Did the hardware configuration change, or did commit ec38c3d9 (nb/amd/amdmct: Select max_lanes based on ECC presence or absence) [1] do that?
Maybe the command below shows you some pointers in the board status repository.
``` diff -ru asus/kgpe-d16/4.3-584-gcc9963e/2016-03-25T19_52_04Z asus/kgpe-d16/4.3-587-gec38c3d/2016-03-26T21_40_56Z/ ```
Thanks,
Paul
[1] https://review.coreboot.org/13725
PS: Here the differences of the time stamps.
``` --- asus/kgpe-d16/4.3-584-gcc9963e/2016-03-25T19_52_04Z/coreboot_timestamps.txt 2016-03-28 10:50:03.751723776 +0200 +++ asus/kgpe-d16/4.3-587-gec38c3d/2016-03-26T21_40_56Z/coreboot_timestamps.txt 2016-03-28 10:50:03.843726364 +0200 @@ -1,21 +1,21 @@ 16 entries total: - 0:1st timestamp 106,862 - 1:start of rom stage 109,067 (2,204) - 2:before ram initialization 703,897 (594,830) - 3:after ram initialization 16,416,242 (15,712,345) - 4:end of romstage 16,507,764 (91,521) - 10:start of ramstage 16,564,607 (56,842) - 30:device enumeration 16,564,610 (3) - 40:device configuration 18,026,405 (1,461,794) - 50:device enable 18,033,787 (7,381) - 60:device initialization 18,035,132 (1,345) - 70:device setup done 18,459,042 (423,910) - 75:cbmem post 18,459,045 (2) - 80:write tables 18,459,046 (1) - 90:load payload 18,470,351 (11,304) - 15:starting LZMA decompress (ignore for x86) 18,470,708 (357) - 16:finished LZMA decompress (ignore for x86) 18,496,547 (25,839) - 99:selfboot jump 18,496,555 (7) + 0:1st timestamp 107,007 + 1:start of rom stage 108,992 (1,984) + 2:before ram initialization 720,604 (611,612) + 3:after ram initialization 17,449,260 (16,728,655) + 4:end of romstage 17,541,247 (91,986) + 10:start of ramstage 17,598,146 (56,899) + 30:device enumeration 17,598,150 (3) + 40:device configuration 19,061,426 (1,463,275) + 50:device enable 19,068,442 (7,015) + 60:device initialization 19,069,861 (1,419) + 70:device setup done 19,494,645 (424,784) + 75:cbmem post 19,494,647 (1) + 80:write tables 19,494,648 (1) + 90:load payload 19,505,952 (11,303) + 15:starting LZMA decompress (ignore for x86) 19,506,309 (357) + 16:finished LZMA decompress (ignore for x86) 19,532,149 (25,839) + 99:selfboot jump 19,532,157 (7) -Total Time: 18,389,685 +Total Time: 19,425,140 ```
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 04/03/2016 01:31 PM, Paul Menzel wrote:
Dear Timothy, dear Damien,
looking through the board status data, I noticed the boot time increased by one second for the ASUS KGPE-D16.
asus/kgpe-d16/4.3-584-gcc9963e/2016-03-25T19_52_04Z/coreboot_timestamps.txt:Total Time: 18,389,685 asus/kgpe-d16/4.3-587-gec38c3d/2016-03-26T21_40_56Z/coreboot_timestamps.txt:Total Time: 19,425,140Did the hardware configuration change, or did commit ec38c3d9 (nb/amd/amdmct: Select max_lanes based on ECC presence or absence) [1] do that?
Maybe the command below shows you some pointers in the board status repository.
diff -ru asus/kgpe-d16/4.3-584-gcc9963e/2016-03-25T19_52_04Z asus/kgpe-d16/4.3-587-gec38c3d/2016-03-26T21_40_56Z/Thanks,
Paul
No, the hardware itself was not changed. The patch was relatively innocuous so I'm surprised the timestamps changed at all, let alone that much.
To make diagnostics easier in the future, perhaps it would make sense to add timestamped stages to the RAM initialization phase? The problem there is that there are multiple nodes starting up so there would need to be a way to add each stage timestamp for a specific node.
- -- Timothy Pearson Raptor Engineering +1 (415) 727-8645 (direct line) +1 (512) 690-0200 (switchboard) http://www.raptorengineeringinc.com
-
Paul Menzel -
Timothy Pearson