Project

General

Profile

Actions

Bug #1276

open

Compile errors made by DEG (rollup)

Added by Evgeny Novikov over 13 years ago. Updated over 13 years ago.

Status:
New
Priority:
Normal
Assignee:
Category:
Environment Generation
Start date:
05/24/2011
Due date:
% Done:

0%

Estimated time:
Detected in build:
svn
Platform:
Published in build:

Description

First of all I want to notice that few such drivers were discovered on investigating #1166 issue (i.e. they are passed to aspectator that reports that it cannot process error_mark tree node). But they also are described in more general terms, Compile error. Some typical error messages are so:

redeclaration of 'var_ahci_sb600_softreset_4_p1' with no linkage
storage size of 'var_group2' isn't known

For further investigation and fixing you can follow this temporary link or perform task with some aspectator model (like 43_1a) on all kernel drivers or upload attached archive.


Files

current--X--drivers---X--defaultlinux-2.6.37--X--43_1a.pax.bz2 (923 KB) current--X--drivers---X--defaultlinux-2.6.37--X--43_1a.pax.bz2 Archive for all drivers of linux 2.6.37 against 43_1a model verification Evgeny Novikov, 05/24/2011 11:27 AM

Related issues 5 (4 open1 closed)

Blocked by Linux Driver Verification - Bug #553: envgen generates incorrect return type for jedec_probeNew11/15/2010

Actions
Blocked by Linux Driver Verification - Bug #1195: drv-env fails to generate ifdefs for interrupt handler in yellowfin.koOpenVadim Mutilin05/11/2011

Actions
Blocked by Linux Driver Verification - Bug #1287: Redeclaration of variables in DEG that lead to compile errorsClosed05/25/2011

Actions
Blocked by Linux Driver Verification - Bug #1293: DEG error: storage size of some variables isn't knownNew05/27/2011

Actions
Blocked by Linux Driver Verification - Bug #1366: Env-gen prints 'no_instrument_function' attribute obtained from function declarations for auxiliary variablesOpenVadim Mutilin06/16/2011

Actions
Actions #1

Updated by Evgeny Novikov over 13 years ago

One of the related issue is #533 (well known jedec_probe problem).

Actions #2

Updated by Vadim Mutilin over 13 years ago

Eugene Novikov wrote:

First of all I want to notice that few such drivers were discovered on investigating #1166 issue (i.e. they are passed to aspectator that reports that it cannot process error_mark tree node). But they also are described in more general terms, Compile error. Some typical error messages are so:
[...]
For further investigation and fixing you can follow this temporary link or perform task with some aspectator model (like 43_1a) on all kernel drivers or upload attached archive.

The issue described here looks like the one which already being fixed in master branch after FAIL of NIGHT tests.
So the link to the run results is obsolete, because the branch model-43_1a was merged with the buggy master and contained the already fixed DEG errors. There is no sense to analyze it. Instead, the run on the newly merged master should be analysed. The pax should already have been generated on cluster (id=224 as far as I remember).

Actions #3

Updated by Evgeny Novikov over 13 years ago

First of all it sounds strange that the issue is fixed since there at least two related bugs that weren't fixed.
Also (may be I'm mistaking) I have launched with non buggy master merged into model-43_1a branch and obtain almost the same problems.

Actions #4

Updated by Vadim Mutilin over 13 years ago

Eugene Novikov wrote:

First of all it sounds strange that the issue is fixed since there at least two related bugs that weren't fixed.

Two bugs attached here are not related to the problem described in the bug (redeclaration and storage size). Also they are not related to the header of the bug, because it contains word OFTEN, but the two attached bugs affect JUST TWO drivers. I suggest to remove dependency of that bugs, bacause they really don't have HIGH priority like this bug (looking to the header only).

Also (may be I'm mistaking) I have launched with non buggy master merged into model-43_1a branch and obtain almost the same problems.

That should be checked, because nobody wants to waste time analysing incorrect runs.

Actions #5

Updated by Evgeny Novikov over 13 years ago

To make it clear I have launched "fixed" model-43_1a model (commit 46733e0).
By the way this bug is a set of all DEG bugs. Some of them are small ones (as blockers). But other bugs should be found out indeed...

Actions #6

Updated by Vadim Mutilin over 13 years ago

Eugene Novikov wrote:

To make it clear I have launched "fixed" model-43_1a model (commit 46733e0).

Thats good.

By the way this bug is a set of all DEG bugs. Some of them are small ones (as blockers). But other bugs should be found out indeed...

I think that fixing all the bugs in DEG including very small ones is not so HIGH priority. (Remembering its regexp nature). The HIGH priority is to find and fix major bugs, not the small ones.

Actions #7

Updated by Pavel Shved over 13 years ago

  • Subject changed from At the moment DEG often generates source code that isn't compilable to At the moment DEG sometimes generates source code that isn't compilable
  • Priority changed from High to Normal

I agree with Vadim. Yes, maingenerator has several issues, and we should, perhaps, list those that concern compilation in one place. However, we're very short on resources, and we should be more careful in estimating our priorities, especially when it concerns small bugs.

Actions #8

Updated by Pavel Shved over 13 years ago

  • Subject changed from At the moment DEG sometimes generates source code that isn't compilable to Compile errors made by DEG (rollup)
Actions

Also available in: Atom PDF