The 16-byte alignment is only needed for SSE2 assembly code, which obviously
doesn't apply on the MIPS. I've checked in the fix to SVN.
--------------------------------------------------
From: "Cyptmon" <[EMAIL PROTECTED]>
Sent: Thursday, October 25, 2007 9:47 AM
To: "Crypto++ Users" <[EMAIL PROTECTED]>
Subject: Re: Build issues
>
> Hi
>
> I had a question, the file config.h defines the following section for
> the executable
>
> #ifndef CRYPTOPP_SECTION_ALIGN16
> #if defined(__GNUC__) && !defined(__APPLE__)
> // the alignment attribute doesn't seem to work
> without this section attribute when -fdata-sections is turned on
> #define CRYPTOPP_SECTION_ALIGN16
> __attribute__((section ("CryptoPP_Align16")))
> #else
> #define CRYPTOPP_SECTION_ALIGN16
> #endif
> #endif
>
>
> Do we need to add the attribute to the executable or is it optional.
> This is where my ELF executable goes out of sync as it defines a
> section CryptoPP::Align16 which is not recognized by the simulator
>
> Thanks
>
>
> On Oct 25, 4:27 am, Cyptmon <[EMAIL PROTECTED]> wrote:
>> Thanks all, I seem to have figured out the problem
>>
>> The ELF executable format is changed using a linker script which needs
>> to be provided at the linking time. It actually works on the ELF
>> generated to move the text and data sections in the executable and if
>> we dont give that script to the linker, it cribs
>>
>> Thanks Wei
>>
>> On Oct 25, 3:52 am, "Wei Dai" <[EMAIL PROTECTED]> wrote:
>>
>> > GCC 3.2 should be ok, as far as I know.
>>
>> > Sorry, but my knowledge of GCC cross compiling and the ELF executable
>> > format
>> > is minimal. I have no idea what is going on here. Maybe the people who
>> > maintain the MIPS simulator have a mailing list that you can post this
>> > question to?
>>
>> > --------------------------------------------------
>> > From: "Cyptmon" <[EMAIL PROTECTED]>
>> > Sent: Wednesday, October 24, 2007 9:32 PM
>> > To: "Crypto++ Users" <[EMAIL PROTECTED]>
>> > Subject: Re: Build issues
>>
>> > > Another thing is I am using mips cross compiler version 3.2 which is
>> > > based on gcc compiler version 3.2. Can this also affect the correct
>> > > working of the library?
>>
>> > > On Oct 24, 11:30 pm, Cyptmon <[EMAIL PROTECTED]> wrote:
>> > >> A minor correction to my earlier post, the elf parsing code that i
>> > >> pasted, it fails in the data section not the code section, however,
>> > >> the checks that fail are the same except that Text is replaced with
>> > >> data
>>
>> > >> Thanks and waiting for your reply
>>
>> > >> On Oct 24, 11:09 pm, Cyptmon <[EMAIL PROTECTED]> wrote:
>>
>> > >> > I tried running a simple program written in C++ and compiling
>> > >> > using
>> > >> > the same command and ran it under the simulator, that works fine,
>> > >> > which tells me that there is something wrong that I am doing when
>> > >> > linking to the library. All other benchmarks compiled with this
>> > >> > cross
>> > >> > compiler run fine as well.
>>
>> > >> > Hope you can give me some insight into the problem
>>
>> > >> > On Oct 24, 6:20 pm, Cyptmon <[EMAIL PROTECTED]> wrote:
>>
>> > >> > > Hi,
>>
>> > >> > > I was just wondering if the compiler options that I use for
>> > >> > > compiling
>> > >> > > the library are fine with the way the library should be
>> > >> > > compiled,
>> > >> > > some compliance issues or something
>>
>> > >> > > I am pasting the options that I use here again for your quick
>> > >> > > reference, in particular the -Wa,-non_shared and -static
>> > >> > > options.
>>
>> > >> > > The error is actually on the simulator, it parses the ELF
>> > >> > > executable
>> > >> > > generated
>>
>> > >> > > stores the address where the executable is loaded in a variable
>> > >> > > named
>> > >> > > "addr"
>>
>> > >> > > It then parses the headers of the executable, in the header
>> > >> > > called
>> > >> > > "__libc_freeres_fn"
>>
>> > >> > > It checks for the following there
>>
>> > >> > > if (strcmp(name, "__libc_freeres_fn") == 0) {
>> > >> > > if (Text_start == 0) {
>> > >> > > Text_seek = offset;
>> > >> > > Text_start = addr;
>> > >> > > Text_size = size;
>> > >> > > } else if (addr != Text_start + Text_size){
>>
>> > >> > > if( (addr - ( Text_start + Text_size)) < 256 ){
>> > >> > > #ifdef DEBUG
>> > >> > > fprintf(stderr,"__libc_freeres_fn Minor header
>> > >> > > aligment
>> > >> > > correction (%x vs %x)\n"
>> > >> > > ,addr,( Text_start + Text_size));
>> > >> > > #endif
>> > >> > > Text_size += size + addr - ( Text_start + Text_size);
>> > >> > > }else
>> > >> > > fatal("__libc_freeres_fn section not contiguous with
>> > >> > > .text
>> > >> > > section\n");
>>
>> > >> > > //The above line is the error that i get
>>
>> > >> > > Here are the flags that I use for compilation
>>
>> > >> > > LDFLAGS
>> > >> > > = -mips2 -mno-abicalls -mabi=32 -mtune=r6000 -msplit-addresses
>> > >> > > -lm -Wa,--start-group -lc -lnss_files -lnss_dns -lresolv -Wa,--end-
>> > >> > > group
>> > >> > > #MLDOPTS +=-mips2 -Wa,-non_shared
>> > >> > > AR = /usr/local/bin/mipseb-linux-ar
>> > >> > > ARFLAGS = -rv
>> > >> > > RANLIB = /usr/local/bin/mipseb-linux-ranlib
>> > >> > > CP = cp
>> > >> > > MKDIR = mkdir
>> > >> > > EGREP = egrep
>> > >> > > UNAME = $(shell uname)
>> > >> > > ISX86 = $(shell uname -m | $(EGREP) -c "i.86|x86|i86")
>> > >> > > ISMINGW = $(shell uname | $(EGREP) -c "MINGW32")
>>
>> > >> > > CXX = /usr/local/bin/mipseb-linux-g++
>> > >> > > CXXFLAGS = -mips2 -mno-abicalls -mabi=32 -mtune=r6000 -msplit-
>> > >> > > addresses -I. -static -Wa,-non_shared -I.
>> > >> > > -I/usr/local/mipseb-linux/
>> > >> > > include/c++/3.2 -I/usr/local/mipseb-linux/include/c++/3.2/backward
>> > >> > > -I/
>> > >> > > usr/local/mipseb-linux/include/c++/3.2/mips-linux-gnu
>>
>> > >> > > #-I. -I/usr/local/mipseb-linux/include/c++/3.2 -I/usr/local/mipseb-
>> > >> > > linux/include/c++/3.2/backward
>> > >> > > -I/usr/local/mipseb-linux/include/c++/
>> > >> > > 3.2/mips-linux-gnu -static -Wa,-non_shared -mips2 -mno-abicalls -
>> > >> > > mabi=32 -mtune=r6000 -msplit-addresses -mips2 -Wa,-non_shared
>>
>> > >> > > On Oct 24, 5:38 pm, "Wei Dai" <[EMAIL PROTECTED]> wrote:
>>
>> > >> > > > Sorry, I have no idea what that means. Google doesn't even
>> > >> > > > have one
>> > >> > > > hit for
>> > >> > > > "data sections not contiguous", which doesn't make any
>> > >> > > > sense... Did
>> > >> > > > you copy
>> > >> > > > the exact error message?
>>
>> > >> > > > --------------------------------------------------
>> > >> > > > From: "Cyptmon" <[EMAIL PROTECTED]>
>> > >> > > > Sent: Tuesday, October 23, 2007 10:57 PM
>> > >> > > > To: "Crypto++ Users" <[EMAIL PROTECTED]>
>> > >> > > > Subject: Re: Build issues
>>
>> > >> > > > > Hi,
>>
>> > >> > > > > I removed some of the socket related #defines as you
>> > >> > > > > suggested
>> > >> > > > > from
>> > >> > > > > config.h
>>
>> > >> > > > > Now the library builds fine
>>
>> > >> > > > > However, when I try to run it on the simulator
>>
>> > >> > > > > It gives the following error for cryptest.exe
>>
>> > >> > > > > ERROR: __libc_subfreeres data sections not contiguous
>>
>> > >> > > > > Would you know anything why this can happen, some
>> > >> > > > > compiler/linker
>> > >> > > > > option causing this behavior, some ordering problem ?
>>
>> > >> > > > > Thanks
>>
>> > >> > > > > On Oct 24, 12:18 am, Cyptmon <[EMAIL PROTECTED]> wrote:
>> > >> > > > >> I will recompile the cross compiler in the night and post
>> > >> > > > >> on the
>> > >> > > > >> list
>> > >> > > > >> if it worked, thanks for the pointers though
>>
>> > >> > > > >> On Oct 23, 11:56 pm, "Wei Dai" <[EMAIL PROTECTED]> wrote:
>>
>> > >> > > > >> > I have a couple of ideas:
>>
>> > >> > > > >> > 1. Edit config.h and comment out the line "#define
>> > >> > > > >> > SOCKETS_AVAILABLE"
>> > >> > > > >> > 2. Recompile your mips glibc with --enable-static-nss
>>
>> > >> > > > >> > --------------------------------------------------
>> > >> > > > >> > From: "Cyptmon" <[EMAIL PROTECTED]>
>> > >> > > > >> > Sent: Tuesday, October 23, 2007 2:09 PM
>> > >> > > > >> > To: "Crypto++ Users" <[EMAIL PROTECTED]>
>> > >> > > > >> > Subject: Build issues
>>
>> > >> > > > >> > > I am trying to build the library for MIPS, I have
>> > >> > > > >> > > changed
>> > >> > > > >> > > the make
>> > >> > > > >> > > file and I am putting the changes to the makefile
>> > >> > > > >> > > towards
>> > >> > > > >> > > the end.
>> > >> > > > >> > > Now
>> > >> > > > >> > > things compile fine but there is a linker error, it
>> > >> > > > >> > > builds
>> > >> > > > >> > > the
>> > >> > > > >> > > library, however, when I try to link it to another
>> > >> > > > >> > > program,
>> > >> > > > >> > > it says,
>> > >> > > > >> > > linking not done...
>>
>> > >> > > > >> > > Here are the changes to the Makefile
>>
>> > >> > > > >> > > LDFLAGS
>> > >> > > > >> > > = -mips2 -mno-abicalls -mabi=32 -mtune=r6000
>> > >> > > > >> > > -msplit-addresses
>> > >> > > > >> > > -lm -Wa,--start-group -lc -lnss_files -lnss_dns -lresolv
>> > >> > > > >> > > -Wa,--end-
>> > >> > > > >> > > group
>> > >> > > > >> > > #MLDOPTS +=-mips2 -Wa,-non_shared
>> > >> > > > >> > > AR = /usr/local/bin/mipseb-linux-ar
>> > >> > > > >> > > ARFLAGS = -rv
>> > >> > > > >> > > RANLIB = /usr/local/bin/mipseb-linux-ranlib
>> > >> > > > >> > > CP = cp
>> > >> > > > >> > > MKDIR = mkdir
>> > >> > > > >> > > EGREP = egrep
>> > >> > > > >> > > UNAME = $(shell uname)
>> > >> > > > >> > > ISX86 = $(shell uname -m | $(EGREP) -c "i.86|x86|i86")
>> > >> > > > >> > > ISMINGW = $(shell uname | $(EGREP) -c "MINGW32")
>>
>> > >> > > > >> > > CXX = /usr/local/bin/mipseb-linux-g++
>> > >> > > > >> > > CXXFLAGS
>> > >> > > > >> > > = -mips2 -mno-abicalls -mabi=32 -mtune=r6000 -msplit-
>> > >> > > > >> > > addresses -I. -static -Wa,-non_shared -I.
>> > >> > > > >> > > -I/usr/local/mipseb-linux/
>> > >> > > > >> > > include/c++/3.2
>> > >> > > > >> > > -I/usr/local/mipseb-linux/include/c++/3.2/backward -I/
>> > >> > > > >> > > usr/local/mipseb-linux/include/c++/3.2/mips-linux-gnu
>>
>> > >> > > > >> > > #-I. -I/usr/local/mipseb-linux/include/c++/3.2
>> > >> > > > >> > > -I/usr/local/mipseb-
>> > >> > > > >> > > linux/include/c++/3.2/backward
>> > >> > > > >> > > -I/usr/local/mipseb-linux/include/c++/
>> > >> > > > >> > > 3.2/mips-linux-gnu -static -Wa,-non_shared -mips2
>> > >> > > > >> > > -mno-abicalls -
>> > >> > > > >> > > mabi=32 -mtune=r6000 -msplit-addresses -mips2
>> > >> > > > >> > > -Wa,-non_shared
>>
>> > >> > > > >> > > SRCS = $(wildcard *.cpp)
>> > >> > > > >> > > ifeq ($(SRCS),) # workaround
>> > >> > > > >> > > wildcard
>> > >> > > > >> > > function
>> > >> > > > >> > > bug in GNU Make 3.77
>> > >> > > > >> > > SRCS = $(shell echo *.cpp)
>> > >> > > > >> > > endif
>> > >> > > > >> > > OBJS = $(SRCS:.cpp=.o)
>> > >> > > > >> > > # test.o needs to be after bench.o for cygwin 1.1.4
>> > >> > > > >> > > (possible ld
>> > >> > > > >> > > bug?)
>> > >> > > > >> > > TESTOBJS = bench.o bench2.o test.o validat1.o
>> > >> > > > >> > > validat2.o
>> > >> > > > >> > > validat3.o
>> > >> > > > >> > > adhoc.o datatest.o regtest.o fipsalgt.o dlltest.o
>> > >> > > > >> > > LIBOBJS = $(filter-out $(TESTOBJS),$(OBJS))
>>
>> > >> > > > >> > > Here is the linker errors that I get
>>
>> > >> > > > >> > > /usr/local/bin/mipseb-linux-g++ -o
>> > >> > > > >> > > cryptest.exe -mips2 -mno-abicalls
>> > >> > > > >> > > -mabi=32 -mtune=r6000 -msplit-addresses -I. -static
>> > >> > > > >> > > -Wa,-non_shared -
>> > >> > > > >> > > I. -I/usr/local/mipseb-linux/include/c++/3.2
>> > >> > > > >> > > -I/usr/local/mipseb-linux/
>> > >> > > > >> > > include/c++/3.2/backward
>> > >> > > > >> > > -I/usr/local/mipseb-linux/include/c++/3.2/
>> > >> > > > >> > > mips-linux-gnu bench.o bench2.o test.o validat1.o
>> > >> > > > >> > > validat2.o
>> > >> > > > >> > > validat3.o adhoc.o datatest.o regtest.o fipsalgt.o
>> > >> > > > >> > > dlltest.o -L. -
>> > >> > > > >> > > lcryptopp -mips2 -mno-abicalls -mabi=32 -mtune=r6000
>> > >> > > > >> > > -msplit-addresses
>> > >> > > > >> > > -lm -Wa,--start-group -lc -lnss_files -lnss_dns
>>
>> ...
>>
>> read more ยป
>
>
> >
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the "Crypto++ Users"
Google Group.
To unsubscribe, send an email to [EMAIL PROTECTED]
More information about Crypto++ and this group is available at
http://www.cryptopp.com.
-~----------~----~----~----~------~----~------~--~---
