This is the mail archive of the
java-patches@gcc.gnu.org
mailing list for the Java project.
Re: Alignment problem with hashtable locks on PowerPC
- To: Jeff Sturm <jsturm at one-point dot com>
- Subject: Re: Alignment problem with hashtable locks on PowerPC
- From: Bryce McKinlay <bryce at waitaki dot otago dot ac dot nz>
- Date: Tue, 13 Nov 2001 11:40:52 +1300
- Cc: "Boehm, Hans" <hans_boehm at hp dot com>, "'java-patches at gcc dot gnu dot org '" <java-patches at gcc dot gnu dot org>
- References: <Pine.LNX.4.10.10111121308200.25507-100000@mars.deadcafe.org>
Jeff Sturm wrote:
>For SPARC I read the thread ID from a global register. Empirically I
>found that %g7 contains thread-specific data, at least on Solaris 7.
>However the ABI reserves this register for system use, so this is probably
>unsuitable for use in libgcj.
>
>There are three other registers reserved for application use. I assume
>libgcj could use any of these for thread-local data? Doesn't the PPC ABI
>have similar application-defined registers?
>
If it did there wouldn't be a problem! I believe the (32-bit) PPC Linux
ABI doesn't reserve any registers for global/application use, ie all of
them are either available to the compiler or used for something already.
There has been talk of changing/fixing the linux ABI, since there are
other "bugs" with it as well, but of course doing that would be a
painful change. Everybody has their own ABI on PPC ;-)
regards
Bryce.