.rdata section in Cygwin executables?
The emacs recompile fails because there is a section in the initially built emacs.exe named .rdata that the unexec() code for Cygwin is not expecting. The section appears to have something to do with exception handling and is only 1k in size. I'm reading up on this at the moment, but a couple questions occur to me that someone here might be able to answer off the top of their head: - can .rdata just be combined with the preceeding .data section? - perhaps there is a way to do this with the linker so I don't have to change the emacs unexec() code? -- Joe Buehler
Re: .rdata section in Cygwin executables?
Joe Buehler wrote: The emacs recompile fails because there is a section in the initially built emacs.exe named .rdata that the unexec() code for Cygwin is not expecting. The section appears to have something to do with exception handling and is only 1k in size. .rdata is a section where read-only data, such as constants, is stored. Most modern OSes and toolchain try to put as much info as possible to read-only segments of binary image because it reduce memory required by programs to run (when program is fork()ed, for instance, read-only pages may be shared between processes) On linux such section is called .rodata, so probably you can take a look how linux version of emacs deals with that. I'm reading up on this at the moment, but a couple questions occur to me that someone here might be able to answer off the top of their head: - can .rdata just be combined with the preceeding .data section? It's possible, but it's not the Right Way (tm). - perhaps there is a way to do this with the linker so I don't have to change the emacs unexec() code? You may write your own linker script so that all rdata section will be put to .data when emacs.exe is linked, but again, it doesn't look as the proper way to deal with the problem. egor.
Re: .rdata section in Cygwin executables?
Egor Duda wrote: You may write your own linker script so that all rdata section will be put to .data when emacs.exe is linked, but again, it doesn't look as the proper way to deal with the problem. According to objdump, there is no readonly flag set for .rdata. Does Windows know about the name .rdata and know that only one copy is needed for multiple emacs processes? Or perhaps Cygwin knows about the name for similar reasons? Or perhaps (gasp) the implementation is not currently complete and .rdata buys me nothing at present? Thanks for the feedback -- I am looking for something easy to do here to get the existing 21.2 emacs back in shape for Cygwin. I already need to rewrite unexec() for the FSF version for copyright reasons and don't really want to spend any time on the current unexec() code -- it's throwaway at some point. -- Joe Buehler
Re: .rdata section in cygwin executables?
On Fri, Mar 12, 2004 at 10:28:02AM -0500, Joe Buehler wrote: Egor Duda wrote: You may write your own linker script so that all rdata section will be put to .data when emacs.exe is linked, but again, it doesn't look as the proper way to deal with the problem. According to objdump, there is no readonly flag set for .rdata. Try a binutils from CVS. There was some funkiness with readonly flags in binutils. I don't know if it infects the cygwin version of binutils or not. I keep rebuilding binutils in preparation for a new release but that effort invariably triggers someone sending an interesting new bug or bugfix to the binutils mailing list so I have been waiting for the code to settle down. cgf
Re: .rdata section in Cygwin executables?
On Friday 12 March 2004 15:51, Egor Duda wrote: Joe Buehler wrote: The emacs recompile fails because there is a section in the initially built emacs.exe named .rdata that the unexec() code for Cygwin is not expecting. The section appears to have something to do with exception handling and is only 1k in size. .rdata is a section where read-only data, such as constants, is stored. Most modern OSes and toolchain try to put as much info as possible to read-only segments of binary image because it reduce memory required by programs to run (when program is fork()ed, for instance, read-only pages may be shared between processes) In case of using gcc 3.3x and using auto-imported vars, you should use the '-fdata-sections' flags for compiling sources because otherwise these variables are not detected as variables. See http://sources.redhat.com/ml/binutils/2003-10/msg00256.html for more infos. Just a note. Ralf
