...hopefully I got the quoting level mess fixed... :-) On Fri, 20 Jul 2007, Li YanBo wrote: > On 7/20/07, Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxx> wrote: > > > > - agnx_read32 & agnx_write32 > > * change to static inline, solves also the use of implicit > > variable (reg) as you have return & assign instead (implicit > > variables confuse readers, see Documentation/CodingStyle) > > It will be fixed later. > > > * You seem to be also doing direct access using ioread/write32 > > here and there?!? > > Yes, do you mean using __raw_readl and __raw_writel are better? ...I was just asking whether you're _not_ using agnx_read32/write32 but ioread32/iowrite32 directly on purpose or not... > > - I'm not sure if the comments per reg in phy_init() are that useful if > > the same information is found in the header file already (just a matter > > of taste though, shouldn't prevent merging, just IMHO) > > That just used to sync with the SPECS, for the SPECS always change, so it help > me to change the code, it will be removed when the SPECS be more stable. Like I indicated already you can as well keep them and nobody is hurt... :-) -- i.
