This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [Patch][BZ 14594] Demangle function pointers before testing them


On 09/21/2012 10:59 AM, Andreas Schwab wrote:
Jeff Law <law@redhat.com> writes:

You're still missing my point.

No, I don't. Your patch is just wrong. I wonder how you tested it.
Most definitely, it is wrong. It built, but I had interrupted the testing process to add another unrelated change. That's my bad.

However, fundamentally the code in glibc as it exists today is wrong as well. It's just wrong in a way that is amazingly hard to test.


Every architecture that currently implements pointer mangling has function addresses aligned to some boundary (even if only by convention).
If it's just convention, then we need to make it a strict policy. I'm fine with that.

Jeff


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]