From: Arne R. <arn...@xi...> - 2006-06-30 12:14:51
|
Am Mittwoch, den 28.06.2006, 18:20 -0400 schrieb Richard Bollinger: > And last but not least, here's the new and improved syncwrite patch > which applies after the Lindent patch. > 1) Moved syncwrite from fileio_data to iet_volume, as I needed access > outside of file-io.c > 2) Added logic to send correct caching_pg data depending on syncwrite setting > > Thats it for me.... time to go bowling! > > Rich B First of all: thanks a lot for your work and persistence on getting this in. After having a glance at it, I have to do some nitpicking ;-), but it might help getting your patch merged: 1. Please send patches inline, it's easier to review (see (3)). 2. Please use separate patches for separate issues, i.e. the IPV6 userland stuff should be rerolled in another patch. I guess that's what Ming tried to express when pointing at that code? 3. in fileio_sync(): + if (++nsyncs > msyncs || res) { + eprintk("res=%d, nsyncs=%u, path=\"%s\"\n", res, nsyncs, + p->path); + if (nsyncs > msyncs) { + msyncs <<= 1; + msyncs |= 1; + } + } Please don't. Error handling (res != 0) and debugging stuff ({n,m}syncs) need to be separated, with the debugging code being optional, i.e. no debugging noise by default. Thanks, Arne -- Arne Redlich Xiranet Communications GmbH |