Bug 124370

Summary: Fix failed to build for Linux/MIPS64EL
Product: WebKit Reporter: YunQiang Su <wzssyqa>
Component: JavaScriptCoreAssignee: Nobody <webkit-unassigned>
Status: RESOLVED FIXED    
Severity: Normal CC: benjamin, berto, cmarcelo, commit-queue, ggaren
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: All   
OS: Linux   
Attachments:
Description Flags
patch for build webkit on Linux/MIPS64EL
none
patch merged into qtwebkit.
none
Patch
none
Patch
none
don't use __mips64 for N64 determination. darin: review+

YunQiang Su
Reported 2013-11-14 10:14:46 PST
There is no process for 64bit mips processors, so for now, it will use JSVAULE32_64 instead of JSCVALUE64 With the attached patch, I can build webkitgtk (2.2.0) and qtwebkit 2.2.1 on Debian Sid.
Attachments
patch for build webkit on Linux/MIPS64EL (2.70 KB, patch)
2013-11-14 10:15 PST, YunQiang Su
no flags
patch merged into qtwebkit. (2.20 KB, patch)
2014-07-10 03:04 PDT, YunQiang Su
no flags
Patch (2.46 KB, patch)
2014-07-10 05:40 PDT, Alberto Garcia
no flags
Patch (2.46 KB, patch)
2015-01-18 11:57 PST, Alberto Garcia
no flags
don't use __mips64 for N64 determination. (2.50 KB, patch)
2015-01-19 01:15 PST, YunQiang Su
darin: review+
YunQiang Su
Comment 1 2013-11-14 10:15:46 PST
Created attachment 216957 [details] patch for build webkit on Linux/MIPS64EL
YunQiang Su
Comment 2 2014-07-10 03:02:34 PDT
This patch has been merged into qtwebkit, and test well. I also test it for webkitgtk. Please consider merge it.
YunQiang Su
Comment 3 2014-07-10 03:04:51 PDT
Created attachment 234697 [details] patch merged into qtwebkit.
Alberto Garcia
Comment 4 2014-07-10 05:40:25 PDT
Created attachment 234705 [details] Patch The patch needs a ChangeLog entry, I just added one. Anyway I think the addition of new architectures must be discussed first in webkit-dev, am I right?
Geoffrey Garen
Comment 5 2014-07-11 14:25:18 PDT
Using JSValue32_64 on a 64bit architecture is pretty weak sauce. (And does it work at all, given that pointers are 64bit?) Is there anyone with enough involvement in the project, who cares about this architecture, who can make it work with JSValue64?
YunQiang Su
Comment 6 2014-07-11 20:01:06 PDT
In this patch, we use JSValue64 not JSValue32_64.
YunQiang Su
Comment 7 2014-10-29 23:46:47 PDT
Any progress of this bug?
Alberto Garcia
Comment 8 2014-11-02 11:21:08 PST
This has been working in Debian for a while now, can anyone review this?
Alberto Garcia
Comment 9 2015-01-18 11:57:20 PST
Created attachment 244865 [details] Patch I've just rebased the patch. Can someone please review?
YunQiang Su
Comment 10 2015-01-19 01:12:15 PST
Sorry a problem discovered #if defined(__mips64) should be defined(_MIPS_SIM_ABI64) && (_MIPS_SIM == _MIPS_SIM_ABI64) as, __mips64 is also defined for N32 abi.
YunQiang Su
Comment 11 2015-01-19 01:15:23 PST
Created attachment 244881 [details] don't use __mips64 for N64 determination.
Alberto Garcia
Comment 12 2015-01-19 01:19:27 PST
Comment on attachment 244881 [details] don't use __mips64 for N64 determination. Ok, thanks
Darin Adler
Comment 13 2015-01-20 09:27:12 PST
Comment on attachment 244881 [details] don't use __mips64 for N64 determination. View in context: https://bugs.webkit.org/attachment.cgi?id=244881&action=review I’m not a platform expert, but this patch looks OK to me. > Source/WTF/wtf/Platform.h:85 > +#if (defined(mips) || defined(__mips__) || defined(MIPS) || defined(_MIPS_) \ > + || defined(__mips64)) This line break just makes things hard to read. We are not trying to fit in 80 columns. Please consider putting this all on one line.
Alberto Garcia
Comment 14 2015-01-20 11:10:58 PST
(In reply to comment #13) > > +#if (defined(mips) || defined(__mips__) || defined(MIPS) || defined(_MIPS_) \ > > + || defined(__mips64)) > > This line break just makes things hard to read. We are not trying to > fit in 80 columns. Please consider putting this all on one line. Sure, I'll do that. Thanks!
Alberto Garcia
Comment 15 2015-01-20 11:12:29 PST
Note You need to log in before you can comment on or make changes to this bug.