Hello Michael,
Thanks for your reply.
I have another question regarding the testing process - what is the set of languages that is acceptable to test with 'make check' on ARM machine?
Also - regarding EEMBC on ARM: has anyone tuned the number of iterations each benchmark should run? Currently, a default number of iterations is used (for example, in iterationsx86.mak file) and from my experiments it seems to produce unstable results.
Thanks, Revital
From: Michael Hope michael.hope@linaro.org To: Revital1 Eres/Haifa/IBM@IBMIL Cc: linaro-toolchain@lists.linaro.org Date: 19/01/2011 02:05 AM Subject: Re: Testing a GCC patch for ARM
On Mon, Jan 10, 2011 at 10:50 PM, Revital1 Eres ERES@il.ibm.com wrote:
Hello,
I have a patch for ARM that I want to test and I'm not sure what's the procedure of testing in Linaro nor to where it should be committed. (GCC trunk? it's currently under bug fixes mode only). I appreciate help with that.
Hi Revital. Sorry for the delay. In general you should either native build or cross build, and then test from there. You can native build on your own hardware or use the boards inside my farm. Once tested, the patch should be sent to gcc-patches@, approved, committed to upstream trunk, then backported to our consolidation branch. This is tricker at the moment due to upstream being in stage 3, so I recommend implementing against trunk, asking hat-in-hand for a pre-review on gcc-patches@, backporting it to our 4.6 branch, and backporting it to our 4.5 branch.
Very brief response I'm afraid. Let me know what you'd like expanded on.
-- Michael