[Bug target/41644] -minimal-toc not helping for toc section exceeding 64k

2017-02-07 Thread msebor at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=41644 Martin Sebor changed: What|Removed |Added Status|UNCONFIRMED |RESOLVED CC|

[Bug target/41644] -minimal-toc not helping for toc section exceeding 64k

2010-11-11 Thread bergner at gcc dot gnu.org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41644 Peter Bergner changed: What|Removed |Added CC||bergner at gcc dot gnu.org --- Comment #1

[Bug target/41644] -minimal-toc not helping for toc section exceeding 64k

2009-10-12 Thread sathishpy at gmail dot com
--- Comment #10 from sathishpy at gmail dot com 2009-10-12 11:02 --- Thanks Jakub, this is very useful information. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41644

[Bug target/41644] -minimal-toc not helping for toc section exceeding 64k

2009-10-12 Thread jakub at gcc dot gnu dot org
--- Comment #9 from jakub at gcc dot gnu dot org 2009-10-12 10:12 --- The documentation is correct. With -mminimal-toc you get just one entry in the .toc per TU. That TU's .toc entry then points into the middle of .toc1 section chunk with TOC stuff for that TU. And your problem is not

[Bug target/41644] -minimal-toc not helping for toc section exceeding 64k

2009-10-12 Thread sathishpy at gmail dot com
--- Comment #8 from sathishpy at gmail dot com 2009-10-12 09:46 --- gcc man page for minimal-toc: >>This option causes GCC to make 'only one TOC entry' for every file. When you specify this option, GCC will produce code that is slower and larger but >>which >>uses extremely littl

[Bug target/41644] -minimal-toc not helping for toc section exceeding 64k

2009-10-12 Thread jakub at gcc dot gnu dot org
--- Comment #7 from jakub at gcc dot gnu dot org 2009-10-12 09:25 --- This is not a bug in -mminimal-toc. It is a limitation of the ppc64 code model. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41644

[Bug target/41644] -minimal-toc not helping for toc section exceeding 64k

2009-10-12 Thread sathishpy at gmail dot com
--- Comment #6 from sathishpy at gmail dot com 2009-10-12 09:17 --- Thanks, I will look at changing our generator to create smaller translation unit. Meanwhile could you please confirm that -mminimal-toc option is broken and will be fixed in one of the future releases? -- http://gcc

[Bug target/41644] -minimal-toc not helping for toc section exceeding 64k

2009-10-09 Thread pinskia at gcc dot gnu dot org
--- Comment #5 from pinskia at gcc dot gnu dot org 2009-10-09 17:13 --- Also the ABIs for powerpc32-linux-gnu and powerpc64-linux-gnu are different by a huge amount. In the 64bit mode, global addresses are made via the TOC register while in 32bit mode, they are done via a simple lis/add

[Bug target/41644] -minimal-toc not helping for toc section exceeding 64k

2009-10-09 Thread jakub at gcc dot gnu dot org
--- Comment #4 from jakub at gcc dot gnu dot org 2009-10-09 12:20 --- For generated code it is obviously best to adjust the generator so that it doesn't generate so big translation units, instead emits more smaller files, or puts all the string literal addresses into an array and uses th

[Bug target/41644] -minimal-toc not helping for toc section exceeding 64k

2009-10-09 Thread sathishpy at gmail dot com
--- Comment #3 from sathishpy at gmail dot com 2009-10-09 11:22 --- I tried the same program with -m32 and it works fine. I have also tried it for 10 times bigger time file (10 string constants and 1 functions) without using -mminimal-toc, so it looks like some issue with gcc onl

[Bug target/41644] -minimal-toc not helping for toc section exceeding 64k

2009-10-09 Thread pinskia at gcc dot gnu dot org
--- Comment #2 from pinskia at gcc dot gnu dot org 2009-10-09 10:54 --- Even though PR 34708 is similar, it is talking about the inliner creating a huge amount of TOC references. The problem here is even more serious. You have a huge TOC references in your source to begin with. --