Phil, On Wed, Sep 26, 2007 at 05:57:41PM +0530, Philip Mucci wrote: > Hi Stefane, > > Greetings from New Delhi. There are other kernel calls that take > vector arguments...writev for example. What is the behavior there if > one of the entries is botched? Does it provide any clue as to which > element was bad? > Don't think they provide any clue.
> From the tool side, it's no problem in user code to 'try and try > again' in such a scenario, like PAPI. > Yes, that's what I am thinking as well. > If this results in an increased likelyhood of kernel adoption, then > I'm all for it. But when all is said and done, if we still get the > big NO, then I'd recommend keeping around all this stuff that got > pulled out. > I'll keep try to simplify to increase our chances but also to make the overall code better. > Does the below require an interface change from the user and/or PFM > side? Please say no. PFM and Perfmon is now going into production in There was no data structure changes. The flags related to these error codes are gone from perfmon.h but if no tool was using them, it should have no impact. The only place I saw this used with in pfmsetup.c the test program that Kevin wrote. > various places (not just SiCortex or PAPI) and it would be nice to > start being sensitive to this issue. For every interface change that > comes out, that's #ifdefs, configure hacking and additional code that > must be maintained. > Good to hear this stuff goes into production somewhere. Enjoy your stay in India. -- -Stephane _______________________________________________ perfmon mailing list [email protected] http://www.hpl.hp.com/hosted/linux/mail-archives/perfmon/
