On Sunday 23 November 2008, Trilok Soni wrote: > > 2) Please use the media/v4l2-i2c-drv.h or media/v4l2-i2c-drv-legacy.h > > header to hide some of the i2c complexity (again, see e.g. saa7115.c). > > The i2c API tends to change a lot (and some changes are upcoming) so What "changes" do you mean? Since this is not a legacy-style driver (yay!), the upcoming changes won't affect it at all. > > using this header will mean that i2c driver changes will be minimal in > > the future. In addition it will ensure that this driver can be compiled > > with older kernels as well once it is part of the v4l-dvb repository. > > I don't agree with having support to compile with older kernels. Right. Folk wanting legacy tvp5146 and tvp5140 support could try to use the legacy drivers from the DaVinci tree. > Even though I2C APIs change as lot it is for good, and creating > abstractions doesn't help as saa7xxx is family of chips where I don't > see the case here. Once this driver is mainlined if someone does i2c > subsystem change which breaks this driver from building then he/she > has to make changes to all the code affecting it. And AFAIK no such change is anticipated. The conversion from legacy style I2C drivers to "new style" driver-model friendly drivers is progressing fairly well, so that legacy support can be completely removed. > I am not in favour of adding support to compile with older kernels. My two cents: I'm not in favor either. In fact that's the general policy for mainline drivers, and I'm surprised to hear any maintainer suggest it be added. - Dave -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
