[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [microblaze-uclinux] GCC toolchain in latest SVN



Thanks John--

It appears that the compiler tools are back to working.
However, gdb is still broken (gives "Memory Fault"),
as it has been for quite some time.     Does the
GDB binary need a similar fix?   (Interestingly,
"mb-linux-gdbtui" does *not* Memory Fault, it
appears to run just fine.)

Thanks,
Steve


John Williams wrote:
Hi Steve,

Steven Hein wrote:
Hi John--

I'm not sure if this is the right place to ask questions
regarding issues with the Petalogix SVN tree.
(Apologies if it is not!).

This morning, I was trying to run 'mb-linux-gdb'
on the Fedora 8 platform where I have been doing my
Petalinux development for quite some time.
(The tools directory was from SVN, probably two
weeks ago).     The 'mb-linux-gdb' tool gave me a
Memory Fault, but all of the other GCC tools
(gcc, etc.) had worked fine.      I updated my SVN
tree this morning, and found that ALL of the glibc
tools give me the Memory Fault on Fedora 8!!
Did something change in the toolchain build?
Would it be possible to build the toolchain on
a box that would provide compatibility with more
platforms?

That's my fault, I broke the tools briefly yesterday but if you do another update all should be well.

For interest, the problem is that when I originally built the tools, it was in a directoy /home/jwilliams/....

alas GCC hard codes the build path into all of the binaries as part of its search path when looking for headers, libraries etc. Now, most sysmtes other than mine don't haev a /home/jwilliams path on them, so it just skips past that and looks in more sensible places like the subtree where the binary is running rfom etc.

However, if you happen to run the tools on a system where there is a /home/jwilliams tree, but you don't have permissions on it, instead of just silently skipping past it, gcc pukes with a permissions error.

So, to work around this I attempted to anonymise the tools by doing a binary replace of all strings jwilliams with xxxxxxxxx, however first tim eI did it I was too hasty and used the wrong number of x's - sigh. instant segfault, and a good lesson in why not to edit binaries.

Next time round I'll build them in a /tmp/random directory.

Sorry for the hassle

John

___________________________
microblaze-uclinux mailing list
microblaze-uclinux@xxxxxxxxxxxxxx
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/


___________________________
microblaze-uclinux mailing list
microblaze-uclinux@xxxxxxxxxxxxxx
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/