On Mon, 25 Mar 2019, Borislav Petkov wrote: > On Sun, Mar 24, 2019 at 05:23:27PM -0700, Paul Walmsley wrote: > > > + if (!IS_ENABLED(CONFIG_EDAC_DEBUG)) > > > + return; > > > > Can all of these debugfs functions be wrapped with an #if ... #endif such > > that, if CONFIG_EDAC_DEBUG is not set, they will all be stripped out by > > the preprocessor? > > Why would you make the code more ugly with ifdeffery? > > Do you have any serious code size constraints so that you absolutely > need to remove a couple of KBs? We'll definitely take the RAM savings that a few #ifdefs will deliver to us. They add up. We're selling chips for embedded use cases, not just big-iron x86 systems. Other EDAC drivers have far more #ifdef lines than the single set that I'm proposing, so I don't understand why you're singling this driver out for criticism. Consider: ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_SDRAM ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_L2C ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_OCRAM ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_OCRAM ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_L2C ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_ETHERNET ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_NAND ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_DMA ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_USB ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_QSPI ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_SDMMC ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_L2C ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_OCRAM ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_ETHERNET ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_NAND ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_DMA ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_USB ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_QSPI ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_SDMMC ./altera_edac.c:#ifdef CONFIG_EDAC_ALTERA_SDRAM ./synopsys_edac.c:#ifdef CONFIG_EDAC_DEBUG ./synopsys_edac.c:#ifdef CONFIG_EDAC_DEBUG ./synopsys_edac.c:#ifdef CONFIG_EDAC_DEBUG ./synopsys_edac.c:#ifdef CONFIG_EDAC_DEBUG ./synopsys_edac.c:#ifdef CONFIG_EDAC_DEBUG - Paul
