[PATCH 0/2] re-work iASL backend

[Colin Ian King]

On 05/06/14 05:11, Alex Hung wrote:
> On 05/29/2014 09:59 PM, Colin King wrote:
>> From: Colin Ian King <colin.king at canonical.com>
>>
>> Currently the iASL interface is not handling extern declarations
>> at all, which is leading to false positives on test such as the
>> syntaxcheck test.  This patchset re-works the interface to the
>> ACPICA assembler/disassembler to allow one to pass over all the
>> tables containing AML so that exernal objects can be resolved
>> correctly.  The patchset includes a re-working of the fwts
>> iasl interface and modifications to the tests that use these.
>>
>> The iasl interface includes some caching of dumped DSDT and SSDT
>> tables, hence the new requirement for a init/deinit phase before using
>> the assembler/disassembler to load and destroy these cached
>> tables.
>>
>> Colin Ian King (2):
>>    lib + tests: re-work iASL backand and API to handle externals (LP:
>>      #1317390)
>>    fwts-test: fix up tests related to iasl and syntaxcheck changes
>>
>>   .../arg-show-tests-0001/arg-show-tests-0001.log    |    2 +-
>>   .../arg-show-tests-full-0001.log                   |    7 +-
>>   fwts-test/disassemble-0001/DSDT.dsl.original       | 6636
>> --------------------
>>   fwts-test/disassemble-0001/DSDT0.dsl.original      | 6636
>> ++++++++++++++++++++
>>   fwts-test/disassemble-0001/SSDT0.dsl.original      |  156 -
>>   fwts-test/disassemble-0001/SSDT1.dsl.original      |  157 +-
>>   fwts-test/disassemble-0001/SSDT2.dsl.original      |  286 +-
>>   fwts-test/disassemble-0001/SSDT3.dsl.original      |  248 +-
>>   fwts-test/disassemble-0001/SSDT4.dsl.original      |  219 +-
>>   fwts-test/disassemble-0001/SSDT5.dsl.original      |  216 +
>>   .../disassemble-0001/disassemble-aml-0001.log      |   12 +-
>>   fwts-test/disassemble-0001/test-0001.sh            |   12 +-
>>   fwts-test/syntaxcheck-0001/syntaxcheck-0001.log    |   84 +-
>>   src/acpi/osilinux/osilinux.c                       |    7 +
>>   src/acpi/syntaxcheck/syntaxcheck.c                 |   46 +-
>>   src/acpica/source/compiler/fwts_iasl_interface.c   |   31 +-
>>   src/acpica/source/compiler/fwts_iasl_interface.h   |    2 +-
>>   src/hpet/hpet_check/hpet_check.c                   |   14 +-
>>   src/lib/include/fwts_iasl.h                        |    8 +-
>>   src/lib/src/fwts_iasl.c                            |  232 +-
>>   20 files changed, 7576 insertions(+), 7435 deletions(-)
>>   delete mode 100644 fwts-test/disassemble-0001/DSDT.dsl.original
>>   create mode 100644 fwts-test/disassemble-0001/DSDT0.dsl.original
>>   delete mode 100644 fwts-test/disassemble-0001/SSDT0.dsl.original
>>   create mode 100644 fwts-test/disassemble-0001/SSDT5.dsl.original
>>
> 
> I tested the patches on two different systems - HP's m4 and Lenovo's
> x200, and the results are quite interesting.
> 
> On the x200, the number of failure decreases from 60 to 59 (see
> x200_results.log), and that's an improvement; however, the number of
> failures in m4 increases from 54 to 72 (see m4_results.log).
> 

The m4 errors increase with the newer version because the older version
of fwts was unable to reassemble the tables and had to exit the
reassembly phase early whereas the newer version was able to complete
the assembly and also find more issues because it did not abort early.
So, it does seem to be a far more improved test.

I can send you the logs if you require Alex.

> Note:
> 1. I applied the patch based on V14.05.00
> (fb5385fec1f3468b479e87c0bba38ab76f60915b)
> 2. I run the test on 14.05 first and then 14.05 with the patches.
> 
> 
> 
> 
> This body part will be downloaded on demand.
>