Message ID | 1459275781-3863-1-git-send-email-robh@kernel.org |
---|---|
State | New |
Headers | show |
On Tuesday 29 March 2016 13:23:00 Rob Herring wrote: > Drivers shouldn't have to care about HAS_IOMEM to compile and having to > causes a Kconfig mess: > > warning: (MEDIA_SUBDRV_AUTOSELECT && VIDEO_CX231XX && INV_MPU6050_I2C) selects I2C_MUX which has unmet direct dependencies (I2C && HAS_IOMEM) > warning: (ST_IRQCHIP && STMMAC_PLATFORM && DWMAC_IPQ806X && DWMAC_LPC18XX && DWMAC_ROCKCHIP && DWMAC_SOCFPGA && DWMAC_STI && TI_CPSW && PINCTRL_ROCKCHIP && PINCTRL_DOVE && POWER_RESET_KEYSTONE && S3C2410_WATCHDOG && VIDEO_OMAP3 && VIDEO_S5P_FIMC && USB_XHCI_MTK && RTC_DRV_AT91SAM9 && LPC18XX_DMAMUX && VIDEO_OMAP4 && HWSPINLOCK_QCOM && ATMEL_ST && QCOM_GSBI && PHY_HI6220_USB) selects MFD_SYSCON which has unmet direct dependencies (HAS_IOMEM) > > Reuse the !MMU variants for !HAS_IOMEM as they are sufficient for our > needs. This fixes build errors for UM allyesconfig: > > drivers/mfd/syscon.c:89:2: error: implicit declaration of function 'iounmap' [-Werror=implicit-function-declaration] > iounmap(base); > > Reported-by: Fengguang Wu <fengguang.wu@intel.com> > Cc: Arnd Bergmann <arnd@arndb.de> > Cc: linux-arch@vger.kernel.org > Signed-off-by: Rob Herring <robh@kernel.org> (adding Richard and the UML list to cc) I actually prototyped a patch that did the opposite: remove the readl/writel/... definitions when HAS_IOMEM is unset. I didn't get far enough to submit it, but see below for what I did. I think it makes sense to do either one or the other. Arnddiff --git a/drivers/char/Kconfig b/drivers/char/Kconfig index a043107..f6dc17a 100644 --- a/drivers/char/Kconfig +++ b/drivers/char/Kconfig @@ -8,6 +8,7 @@ source "drivers/tty/Kconfig" config DEVMEM bool "/dev/mem virtual device support" + depends on HAS_IOMEM default y help Say Y here if you want to support the /dev/mem device. @@ -17,6 +18,7 @@ config DEVMEM config DEVKMEM bool "/dev/kmem virtual device support" + depends on HAS_IOMEM default y help Say Y here if you want to support the /dev/kmem device. The @@ -94,6 +96,7 @@ config BFIN_OTP_WRITE_ENABLE config PRINTER tristate "Parallel printer support" depends on PARPORT + depends on HAS_IOPORT ---help--- If you intend to attach a printer to the parallel port of your Linux box (as opposed to using a serial printer; if the connector at the diff --git a/drivers/char/mem.c b/drivers/char/mem.c index 4f6f94c..eedd129 100644 --- a/drivers/char/mem.c +++ b/drivers/char/mem.c @@ -91,6 +91,7 @@ void __weak unxlate_dev_mem_ptr(phys_addr_t phys, void *addr) } #endif +#ifdef CONFIG_DEVMEM /* * This funcion reads the *physical* memory. The f_pos points directly to the * memory location. @@ -219,6 +220,7 @@ static ssize_t write_mem(struct file *file, const char __user *buf, *ppos += written; return written; } +#endif int __weak phys_mem_access_prot_allowed(struct file *file, unsigned long pfn, unsigned long size, pgprot_t *vma_prot) @@ -312,6 +314,7 @@ static inline int private_mapping_ok(struct vm_area_struct *vma) } #endif +#ifdef CONFIG_DEVMEM static const struct vm_operations_struct mmap_mem_ops = { #ifdef CONFIG_HAVE_IOREMAP_PROT .access = generic_access_phys @@ -351,7 +354,9 @@ static int mmap_mem(struct file *file, struct vm_area_struct *vma) } return 0; } +#endif +#ifdef CONFIG_DEVKMEM static int mmap_kmem(struct file *file, struct vm_area_struct *vma) { unsigned long pfn; @@ -552,7 +557,9 @@ static ssize_t write_kmem(struct file *file, const char __user *buf, *ppos = p; return virtr + wrote ? : err; } +#endif +#ifdef CONFIG_DEVPORT static ssize_t read_port(struct file *file, char __user *buf, size_t count, loff_t *ppos) { @@ -594,6 +601,7 @@ static ssize_t write_port(struct file *file, const char __user *buf, *ppos = i; return tmp-buf; } +#endif static ssize_t read_null(struct file *file, char __user *buf, size_t count, loff_t *ppos) @@ -710,10 +718,12 @@ static loff_t memory_lseek(struct file *file, loff_t offset, int orig) return ret; } +#ifdef CONFIG_DEVPORT static int open_port(struct inode *inode, struct file *filp) { return capable(CAP_SYS_RAWIO) ? 0 : -EPERM; } +#endif #define zero_lseek null_lseek #define full_lseek null_lseek @@ -722,6 +732,7 @@ static int open_port(struct inode *inode, struct file *filp) #define open_mem open_port #define open_kmem open_mem +#ifdef CONFIG_DEVMEM static const struct file_operations __maybe_unused mem_fops = { .llseek = memory_lseek, .read = read_mem, @@ -733,7 +744,9 @@ static const struct file_operations __maybe_unused mem_fops = { .mmap_capabilities = memory_mmap_capabilities, #endif }; +#endif +#ifdef CONFIG_DEVKMEM static const struct file_operations __maybe_unused kmem_fops = { .llseek = memory_lseek, .read = read_kmem, @@ -745,6 +758,7 @@ static const struct file_operations __maybe_unused kmem_fops = { .mmap_capabilities = memory_mmap_capabilities, #endif }; +#endif static const struct file_operations null_fops = { .llseek = null_lseek, @@ -755,12 +769,14 @@ static const struct file_operations null_fops = { .splice_write = splice_write_null, }; +#ifdef CONFIG_DEVPORT static const struct file_operations __maybe_unused port_fops = { .llseek = memory_lseek, .read = read_port, .write = write_port, .open = open_port, }; +#endif static const struct file_operations zero_fops = { .llseek = zero_lseek, diff --git a/drivers/clocksource/Kconfig b/drivers/clocksource/Kconfig index 33db740..6b58974 100644 --- a/drivers/clocksource/Kconfig +++ b/drivers/clocksource/Kconfig @@ -38,6 +38,7 @@ config DIGICOLOR_TIMER config DW_APB_TIMER bool "DW APB timer driver" if COMPILE_TEST depends on GENERIC_CLOCKEVENTS + depends on HAS_IOMEM help Enables the support for the dw_apb timer. @@ -64,6 +65,7 @@ config ARMADA_370_XP_TIMER config MESON6_TIMER bool "Meson6 timer driver" if COMPILE_TEST depends on GENERIC_CLOCKEVENTS + depends on HAS_IOMEM select CLKSRC_MMIO help Enables the support for the Meson6 timer driver. @@ -114,6 +116,7 @@ config CADENCE_TTC_TIMER config ASM9260_TIMER bool "ASM9260 timer driver" if COMPILE_TEST depends on GENERIC_CLOCKEVENTS + depends on HAS_IOMEM select CLKSRC_MMIO select CLKSRC_OF help diff --git a/drivers/fmc/Kconfig b/drivers/fmc/Kconfig index 3a75f42..adf1391 100644 --- a/drivers/fmc/Kconfig +++ b/drivers/fmc/Kconfig @@ -4,6 +4,7 @@ menuconfig FMC tristate "FMC support" + depends on HAS_IOMEM help FMC (FPGA Mezzanine Carrier) is a mechanical and electrical diff --git a/drivers/fpga/Kconfig b/drivers/fpga/Kconfig index c9b9fdf..36c54af 100644 --- a/drivers/fpga/Kconfig +++ b/drivers/fpga/Kconfig @@ -21,6 +21,7 @@ config FPGA_MGR_SOCFPGA config FPGA_MGR_ZYNQ_FPGA tristate "Xilinx Zynq FPGA" + depends on HAS_IOMEM help FPGA manager driver support for Xilinx Zynq FPGAs. diff --git a/drivers/hwtracing/intel_th/Kconfig b/drivers/hwtracing/intel_th/Kconfig index b7a9073..467dae9 100644 --- a/drivers/hwtracing/intel_th/Kconfig +++ b/drivers/hwtracing/intel_th/Kconfig @@ -1,5 +1,6 @@ config INTEL_TH tristate "Intel(R) Trace Hub controller" + depends on HAS_IOMEM help Intel(R) Trace Hub (TH) is a set of hardware blocks (subdevices) that produce, switch and output trace data from multiple hardware and diff --git a/drivers/mfd/Kconfig b/drivers/mfd/Kconfig index 9ca66de..6ff6246 100644 --- a/drivers/mfd/Kconfig +++ b/drivers/mfd/Kconfig @@ -999,6 +999,7 @@ config MFD_SUN6I_PRCM config MFD_SYSCON bool "System Controller Register R/W Based on Regmap" + depends on HAS_IOMEM select REGMAP_MMIO help Select this option to enable accessing system control registers diff --git a/drivers/misc/altera-stapl/Kconfig b/drivers/misc/altera-stapl/Kconfig index 7f01d8e..c7e4c77 100644 --- a/drivers/misc/altera-stapl/Kconfig +++ b/drivers/misc/altera-stapl/Kconfig @@ -2,7 +2,7 @@ comment "Altera FPGA firmware download module" config ALTERA_STAPL tristate "Altera FPGA firmware download module" - depends on I2C + depends on I2C && HAS_IOPORT default n help An Altera FPGA module. Say Y when you want to support this tool. diff --git a/drivers/mtd/chips/Kconfig b/drivers/mtd/chips/Kconfig index 3b3dabc..189ddce 100644 --- a/drivers/mtd/chips/Kconfig +++ b/drivers/mtd/chips/Kconfig @@ -3,6 +3,7 @@ menu "RAM/ROM/Flash chip drivers" config MTD_CFI tristate "Detect flash chips by Common Flash Interface (CFI) probe" + depends on HAS_IOMEM select MTD_GEN_PROBE select MTD_CFI_UTIL help @@ -15,6 +16,7 @@ config MTD_CFI config MTD_JEDECPROBE tristate "Detect non-CFI AMD/JEDEC-compatible flash chips" + depends on HAS_IOMEM select MTD_GEN_PROBE select MTD_CFI_UTIL help @@ -207,12 +209,14 @@ config MTD_CFI_UTIL config MTD_RAM tristate "Support for RAM chips in bus mapping" + depends on HAS_IOMEM help This option enables basic support for RAM chips accessed through a bus mapping driver. config MTD_ROM tristate "Support for ROM chips in bus mapping" + depends on HAS_IOMEM help This option enables basic support for ROM chips accessed through a bus mapping driver. diff --git a/drivers/mtd/lpddr/Kconfig b/drivers/mtd/lpddr/Kconfig index 3a19cbe..07dd95a 100644 --- a/drivers/mtd/lpddr/Kconfig +++ b/drivers/mtd/lpddr/Kconfig @@ -4,6 +4,7 @@ menu "LPDDR & LPDDR2 PCM memory drivers" config MTD_LPDDR tristate "Support for LPDDR flash chips" select MTD_QINFO_PROBE + depends on HAS_IOMEM help This option enables support of LPDDR (Low power double data rate) flash chips. Synonymous with Mobile-DDR. It is a new standard for diff --git a/drivers/mtd/maps/Kconfig b/drivers/mtd/maps/Kconfig index 7c95a65..86a9604 100644 --- a/drivers/mtd/maps/Kconfig +++ b/drivers/mtd/maps/Kconfig @@ -3,7 +3,8 @@ menu "Mapping drivers for chip access" depends on HAS_IOMEM config MTD_COMPLEX_MAPPINGS - bool "Support non-linear mappings of flash chips" + bool "Support non-linear mappings of flash chips" if HAS_IOMEM + default !HAS_IOMEM help This causes the chip drivers to allow for complicated paged mappings of flash chips. diff --git a/drivers/mtd/nand/Kconfig b/drivers/mtd/nand/Kconfig index 20f01b3..6ed3de7 100644 --- a/drivers/mtd/nand/Kconfig +++ b/drivers/mtd/nand/Kconfig @@ -12,7 +12,7 @@ config MTD_NAND_ECC_SMC menuconfig MTD_NAND tristate "NAND Device Support" - depends on MTD + depends on MTD && HAS_IOMEM select MTD_NAND_IDS select MTD_NAND_ECC help diff --git a/drivers/mtd/spi-nor/Kconfig b/drivers/mtd/spi-nor/Kconfig index 0dc9275..83befab 100644 --- a/drivers/mtd/spi-nor/Kconfig +++ b/drivers/mtd/spi-nor/Kconfig @@ -9,6 +9,7 @@ if MTD_SPI_NOR config MTD_MT81xx_NOR tristate "Mediatek MT81xx SPI NOR flash controller" + depends on HAS_IOMEM help This enables access to SPI NOR flash, using MT81xx SPI NOR flash controller. This controller does not support generic SPI BUS, it only diff --git a/drivers/net/can/Kconfig b/drivers/net/can/Kconfig index 6d04183..b1e268b 100644 --- a/drivers/net/can/Kconfig +++ b/drivers/net/can/Kconfig @@ -132,6 +132,7 @@ config CAN_RCAR config CAN_SUN4I tristate "Allwinner A10 CAN controller" depends on MACH_SUN4I || MACH_SUN7I || COMPILE_TEST + depends on HAS_IOMEM ---help--- Say Y here if you want to use CAN controller found on Allwinner A10/A20 SoCs. diff --git a/drivers/net/hamradio/Kconfig b/drivers/net/hamradio/Kconfig index bf5e596..507f7a5 100644 --- a/drivers/net/hamradio/Kconfig +++ b/drivers/net/hamradio/Kconfig @@ -113,7 +113,8 @@ config SCC_TRXECHO config BAYCOM_SER_FDX tristate "BAYCOM ser12 fullduplex driver for AX.25" - depends on AX25 && !S390 + depends on AX25 + depends on HAS_IOPORT select CRC_CCITT ---help--- This is one of two drivers for Baycom style simple amateur radio @@ -133,7 +134,8 @@ config BAYCOM_SER_FDX config BAYCOM_SER_HDX tristate "BAYCOM ser12 halfduplex driver for AX.25" - depends on AX25 && !S390 + depends on AX25 + depends on HAS_IOPORT select CRC_CCITT ---help--- This is one of two drivers for Baycom style simple amateur radio @@ -181,7 +183,8 @@ config BAYCOM_EPP config YAM tristate "YAM driver for AX.25" - depends on AX25 && !S390 + depends on AX25 + depends on HAS_IOPORT help The YAM is a modem for packet radio which connects to the serial port and includes some of the functions of a Terminal Node diff --git a/drivers/nvmem/Kconfig b/drivers/nvmem/Kconfig index bc4ea58..6ef3d0a 100644 --- a/drivers/nvmem/Kconfig +++ b/drivers/nvmem/Kconfig @@ -1,6 +1,7 @@ menuconfig NVMEM tristate "NVMEM Support" select REGMAP + depends on HAS_IOMEM help Support for NVMEM(Non Volatile Memory) devices like EEPROM, EFUSES... diff --git a/drivers/phy/Kconfig b/drivers/phy/Kconfig index e7e117d..d2216df 100644 --- a/drivers/phy/Kconfig +++ b/drivers/phy/Kconfig @@ -30,9 +30,10 @@ config PHY_BERLIN_SATA Enable this to support the SATA PHY on Marvell Berlin SoCs. config ARMADA375_USBCLUSTER_PHY - def_bool y - depends on MACH_ARMADA_375 || COMPILE_TEST + bool "ARMADA 375 USB cluster PHY driver" if COMPILE_TEST && !MACH_ARMADA_375 + def_bool MACH_ARMADA_375 depends on OF + depends on HAS_IOMEM select GENERIC_PHY config PHY_DM816X_USB @@ -128,6 +129,7 @@ config PHY_RCAR_GEN3_USB2 config OMAP_CONTROL_PHY tristate "OMAP CONTROL PHY Driver" depends on ARCH_OMAP2PLUS || COMPILE_TEST + depends on HAS_IOMEM help Enable this to add support for the PHY part present in the control module. This driver has API to power on the USB2 PHY and to write to @@ -224,6 +226,8 @@ config PHY_MT65XX_USB3 config PHY_HI6220_USB tristate "hi6220 USB PHY support" + depends on ARM64 || COMPILE_TEST + depends on HAS_IOMEM select GENERIC_PHY select MFD_SYSCON help @@ -401,6 +405,7 @@ config PHY_CYGNUS_PCIE tristate "Broadcom Cygnus PCIe PHY driver" depends on OF && (ARCH_BCM_CYGNUS || COMPILE_TEST) select GENERIC_PHY + depends on HAS_IOMEM default ARCH_BCM_CYGNUS help Enable this to support the Broadcom Cygnus PCIe PHY. diff --git a/drivers/power/reset/Kconfig b/drivers/power/reset/Kconfig index 1131cf7..bb5bd1b 100644 --- a/drivers/power/reset/Kconfig +++ b/drivers/power/reset/Kconfig @@ -41,7 +41,7 @@ config POWER_RESET_AXXIA config POWER_RESET_BRCMSTB bool "Broadcom STB reset driver" depends on ARM || MIPS || COMPILE_TEST - depends on MFD_SYSCON + depends on MFD_SYSCON && HAS_IOMEM default ARCH_BRCMSTB help This driver provides restart support for Broadcom STB boards. @@ -148,6 +148,7 @@ config POWER_RESET_KEYSTONE config POWER_RESET_SYSCON bool "Generic SYSCON regmap reset driver" depends on OF + depends on HAS_IOMEM select MFD_SYSCON help Reboot support for generic SYSCON mapped register reset. @@ -155,6 +156,7 @@ config POWER_RESET_SYSCON config POWER_RESET_SYSCON_POWEROFF bool "Generic SYSCON regmap poweroff driver" depends on OF + depends on HAS_IOMEM select MFD_SYSCON help Poweroff support for generic SYSCON mapped register poweroff. diff --git a/drivers/staging/comedi/Kconfig b/drivers/staging/comedi/Kconfig index e7255f8..e1ba4aa 100644 --- a/drivers/staging/comedi/Kconfig +++ b/drivers/staging/comedi/Kconfig @@ -67,6 +67,7 @@ config COMEDI_TEST config COMEDI_PARPORT tristate "Parallel port support" + depends on HAS_IOPORT ---help--- Enable support for the standard parallel port. A cheap and easy way to get a few more digital I/O lines. Steal @@ -87,6 +88,7 @@ config COMEDI_SERIAL2002 config COMEDI_SSV_DNP tristate "SSV Embedded Systems DIL/Net-PC support" depends on X86_32 || COMPILE_TEST + depends on HAS_IOPORT ---help--- Enable support for SSV Embedded Systems DIL/Net-PC @@ -97,6 +99,7 @@ endif # COMEDI_MISC_DRIVERS menuconfig COMEDI_ISA_DRIVERS bool "Comedi ISA and PC/104 drivers" + depends on HAS_IOPORT ---help--- Enable comedi ISA and PC/104 drivers to be built @@ -1110,7 +1113,7 @@ endif # COMEDI_PCI_DRIVERS menuconfig COMEDI_PCMCIA_DRIVERS tristate "Comedi PCMCIA drivers" - depends on PCMCIA + depends on PCMCIA && HAS_IOPORT ---help--- Enable support for comedi PCMCIA drivers. @@ -1262,6 +1265,7 @@ config COMEDI_8255 config COMEDI_8255_SA tristate "Standalone 8255 support" select COMEDI_8255 + depends on HAS_IOPORT ---help--- Enable support for 8255 digital I/O as a standalone driver. diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig index 8cc4ac6..31ae027 100644 --- a/drivers/thermal/Kconfig +++ b/drivers/thermal/Kconfig @@ -178,6 +178,7 @@ config THERMAL_EMULATION config HISI_THERMAL tristate "Hisilicon thermal driver" depends on (ARCH_HISI && CPU_THERMAL && OF) || COMPILE_TEST + depends on HAS_IOMEM help Enable this to plug hisilicon's thermal sensor driver into the Linux thermal framework. cpufreq is used as the cooling device to throttle @@ -198,6 +199,7 @@ config SPEAR_THERMAL bool "SPEAr thermal sensor driver" depends on PLAT_SPEAR || COMPILE_TEST depends on OF + depends on HAS_IOMEM help Enable this to plug the SPEAr thermal sensor driver into the Linux thermal framework. @@ -206,6 +208,7 @@ config ROCKCHIP_THERMAL tristate "Rockchip thermal driver" depends on ARCH_ROCKCHIP || COMPILE_TEST depends on RESET_CONTROLLER + depends on HAS_IOMEM help Rockchip thermal driver provides support for Temperature sensor ADC (TS-ADC) found on Rockchip SoCs. It supports one critical @@ -216,6 +219,7 @@ config RCAR_THERMAL tristate "Renesas R-Car thermal driver" depends on ARCH_SHMOBILE || COMPILE_TEST depends on HAS_IOMEM + depends on HAS_IOMEM help Enable this to plug the R-Car thermal sensor driver into the Linux thermal framework. @@ -223,6 +227,7 @@ config RCAR_THERMAL config KIRKWOOD_THERMAL tristate "Temperature sensor on Marvell Kirkwood SoCs" depends on MACH_KIRKWOOD || COMPILE_TEST + depends on HAS_IOMEM depends on OF help Support for the Kirkwood thermal sensor driver into the Linux thermal @@ -231,6 +236,7 @@ config KIRKWOOD_THERMAL config DOVE_THERMAL tristate "Temperature sensor on Marvell Dove SoCs" depends on ARCH_DOVE || MACH_DOVE || COMPILE_TEST + depends on HAS_IOMEM depends on OF help Support for the Dove thermal sensor driver in the Linux thermal @@ -249,6 +255,7 @@ config DB8500_THERMAL config ARMADA_THERMAL tristate "Armada 370/XP thermal management" depends on ARCH_MVEBU || COMPILE_TEST + depends on HAS_IOMEM depends on OF help Enable this option if you want to have support for thermal management @@ -366,12 +373,12 @@ config INTEL_PCH_THERMAL programmable trip points and other information. menu "Texas Instruments thermal drivers" -depends on ARCH_HAS_BANDGAP || COMPILE_TEST +depends on ARCH_HAS_BANDGAP || (COMPILE_TEST && HAS_IOMEM) source "drivers/thermal/ti-soc-thermal/Kconfig" endmenu menu "Samsung thermal drivers" -depends on ARCH_EXYNOS || COMPILE_TEST +depends on ARCH_EXYNOS || (COMPILE_TEST && HAS_IOMEM) source "drivers/thermal/samsung/Kconfig" endmenu diff --git a/include/linux/irq.h b/include/linux/irq.h index 3c1c967..a5d0f12 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -916,6 +916,7 @@ static inline void irq_gc_lock(struct irq_chip_generic *gc) { } static inline void irq_gc_unlock(struct irq_chip_generic *gc) { } #endif +#ifdef CONFIG_HAS_IOMEM static inline void irq_reg_writel(struct irq_chip_generic *gc, u32 val, int reg_offset) { @@ -933,5 +934,6 @@ static inline u32 irq_reg_readl(struct irq_chip_generic *gc, else return readl(gc->reg_base + reg_offset); } +#endif #endif /* _LINUX_IRQ_H */ diff --git a/include/linux/mtd/cfi.h b/include/linux/mtd/cfi.h index 9b57a9b..d5db153 100644 --- a/include/linux/mtd/cfi.h +++ b/include/linux/mtd/cfi.h @@ -310,6 +310,7 @@ uint32_t cfi_send_gen_cmd(u_char cmd, uint32_t cmd_addr, uint32_t base, struct map_info *map, struct cfi_private *cfi, int type, map_word *prev_val); +#ifdef CONFIG_HAS_IOMEM static inline uint8_t cfi_read_query(struct map_info *map, uint32_t addr) { map_word val = map_read(map, addr); @@ -341,6 +342,7 @@ static inline uint16_t cfi_read_query16(struct map_info *map, uint32_t addr) return cfi32_to_cpu(map, val.x[0]); } } +#endif void cfi_udelay(int us); diff --git a/include/linux/mtd/map.h b/include/linux/mtd/map.h index 58f3ba7..07c0d497 100644 --- a/include/linux/mtd/map.h +++ b/include/linux/mtd/map.h @@ -410,6 +410,7 @@ static inline map_word map_word_ff(struct map_info *map) return r; } +#ifdef CONFIG_HAS_IOMEM static inline map_word inline_map_read(struct map_info *map, unsigned long ofs) { map_word r; @@ -463,6 +464,7 @@ static inline void inline_map_copy_to(struct map_info *map, unsigned long to, co { memcpy_toio(map->virt + to, from, len); } +#endif #ifdef CONFIG_MTD_COMPLEX_MAPPINGS #define map_read(map, ofs) (map)->read(map, ofs) diff --git a/kernel/resource.c b/kernel/resource.c index 09c0597..efd5c9e 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -29,7 +29,9 @@ struct resource ioport_resource = { .name = "PCI IO", .start = 0, +#ifdef CONFIG_HAS_IOPORT .end = IO_SPACE_LIMIT, +#endif .flags = IORESOURCE_IO, }; EXPORT_SYMBOL(ioport_resource); diff --git a/lib/Kconfig b/lib/Kconfig index 133ebc0..38ff36d 100644 --- a/lib/Kconfig +++ b/lib/Kconfig @@ -382,9 +382,14 @@ config HAS_IOMEM select GENERIC_IO default y +config HAS_IOPORT + bool + depends on HAS_IOMEM && !NO_IOPORT + default y + config HAS_IOPORT_MAP bool - depends on HAS_IOMEM && !NO_IOPORT_MAP + depends on HAS_IOMEM && HAS_IOPORT && !NO_IOPORT_MAP default y config HAS_DMA diff --git a/mm/bootmem.c b/mm/bootmem.c index 91e32bc..3471237 100644 --- a/mm/bootmem.c +++ b/mm/bootmem.c @@ -99,7 +99,7 @@ static unsigned long __init init_bootmem_core(bootmem_data_t *bdata, unsigned long mapsize; mminit_validate_memmodel_limits(&start, &end); - bdata->node_bootmem_map = phys_to_virt(PFN_PHYS(mapstart)); + bdata->node_bootmem_map = __va(PFN_PHYS(mapstart)); bdata->node_min_pfn = start; bdata->node_low_pfn = end; link_bootmem(bdata); @@ -585,7 +585,7 @@ find_block: PFN_UP(end_off), BOOTMEM_EXCLUSIVE)) BUG(); - region = phys_to_virt(PFN_PHYS(bdata->node_min_pfn) + + region = __va(PFN_PHYS(bdata->node_min_pfn) + start_off); memset(region, 0, size); /*
Am 29.03.2016 um 21:37 schrieb Arnd Bergmann: > On Tuesday 29 March 2016 13:23:00 Rob Herring wrote: >> Drivers shouldn't have to care about HAS_IOMEM to compile and having to >> causes a Kconfig mess: >> >> warning: (MEDIA_SUBDRV_AUTOSELECT && VIDEO_CX231XX && INV_MPU6050_I2C) selects I2C_MUX which has unmet direct dependencies (I2C && HAS_IOMEM) >> warning: (ST_IRQCHIP && STMMAC_PLATFORM && DWMAC_IPQ806X && DWMAC_LPC18XX && DWMAC_ROCKCHIP && DWMAC_SOCFPGA && DWMAC_STI && TI_CPSW && PINCTRL_ROCKCHIP && PINCTRL_DOVE && POWER_RESET_KEYSTONE && S3C2410_WATCHDOG && VIDEO_OMAP3 && VIDEO_S5P_FIMC && USB_XHCI_MTK && RTC_DRV_AT91SAM9 && LPC18XX_DMAMUX && VIDEO_OMAP4 && HWSPINLOCK_QCOM && ATMEL_ST && QCOM_GSBI && PHY_HI6220_USB) selects MFD_SYSCON which has unmet direct dependencies (HAS_IOMEM) >> >> Reuse the !MMU variants for !HAS_IOMEM as they are sufficient for our >> needs. This fixes build errors for UM allyesconfig: >> >> drivers/mfd/syscon.c:89:2: error: implicit declaration of function 'iounmap' [-Werror=implicit-function-declaration] >> iounmap(base); >> >> Reported-by: Fengguang Wu <fengguang.wu@intel.com> >> Cc: Arnd Bergmann <arnd@arndb.de> >> Cc: linux-arch@vger.kernel.org >> Signed-off-by: Rob Herring <robh@kernel.org> > > (adding Richard and the UML list to cc) For the above error I've sent already a fix: https://lkml.org/lkml/2016/3/25/321 > I actually prototyped a patch that did the opposite: remove the readl/writel/... > definitions when HAS_IOMEM is unset. I didn't get far enough to submit it, > but see below for what I did. > > I think it makes sense to do either one or the other. > > Arnd > > diff --git a/drivers/char/Kconfig b/drivers/char/Kconfig > index a043107..f6dc17a 100644 > --- a/drivers/char/Kconfig > +++ b/drivers/char/Kconfig > @@ -8,6 +8,7 @@ source "drivers/tty/Kconfig" > > config DEVMEM > bool "/dev/mem virtual device support" > + depends on HAS_IOMEM Hmm, this means no /dev/mem device for UML? (And some s390 variants). Not sure if this a good idea. But I like it more than having ioremap/iounmap stubs for !HAS_IOMEM. :) Thanks, //richard
On Tue, Mar 29, 2016 at 2:37 PM, Arnd Bergmann <arnd@arndb.de> wrote: > On Tuesday 29 March 2016 13:23:00 Rob Herring wrote: >> Drivers shouldn't have to care about HAS_IOMEM to compile and having to >> causes a Kconfig mess: >> >> warning: (MEDIA_SUBDRV_AUTOSELECT && VIDEO_CX231XX && INV_MPU6050_I2C) selects I2C_MUX which has unmet direct dependencies (I2C && HAS_IOMEM) >> warning: (ST_IRQCHIP && STMMAC_PLATFORM && DWMAC_IPQ806X && DWMAC_LPC18XX && DWMAC_ROCKCHIP && DWMAC_SOCFPGA && DWMAC_STI && TI_CPSW && PINCTRL_ROCKCHIP && PINCTRL_DOVE && POWER_RESET_KEYSTONE && S3C2410_WATCHDOG && VIDEO_OMAP3 && VIDEO_S5P_FIMC && USB_XHCI_MTK && RTC_DRV_AT91SAM9 && LPC18XX_DMAMUX && VIDEO_OMAP4 && HWSPINLOCK_QCOM && ATMEL_ST && QCOM_GSBI && PHY_HI6220_USB) selects MFD_SYSCON which has unmet direct dependencies (HAS_IOMEM) >> >> Reuse the !MMU variants for !HAS_IOMEM as they are sufficient for our >> needs. This fixes build errors for UM allyesconfig: >> >> drivers/mfd/syscon.c:89:2: error: implicit declaration of function 'iounmap' [-Werror=implicit-function-declaration] >> iounmap(base); >> >> Reported-by: Fengguang Wu <fengguang.wu@intel.com> >> Cc: Arnd Bergmann <arnd@arndb.de> >> Cc: linux-arch@vger.kernel.org >> Signed-off-by: Rob Herring <robh@kernel.org> > > (adding Richard and the UML list to cc) > > I actually prototyped a patch that did the opposite: remove the readl/writel/... > definitions when HAS_IOMEM is unset. I didn't get far enough to submit it, > but see below for what I did. Ewww. Why do the opposite of what we do for every other Kconfig symbol which is provide empty functions? It really only functions as a disable on UML flag which runs counter to enabling drivers to build for all arches. I actually started a patch to remove the HAS_IOMEM dependency everywhere (or just the per driver cases). It didn't break as bad as I expected, but became more than I wanted to fix. Mainly, all the devm_ variants also need empty versions or to be always enabled. Rob
On Wed, Mar 30, 2016 at 3:23 AM, Rob Herring <robh@kernel.org> wrote: > Drivers shouldn't have to care about HAS_IOMEM to compile and having to > causes a Kconfig mess: > > warning: (MEDIA_SUBDRV_AUTOSELECT && VIDEO_CX231XX && INV_MPU6050_I2C) selects I2C_MUX which has unmet direct dependencies (I2C && HAS_IOMEM) > warning: (ST_IRQCHIP && STMMAC_PLATFORM && DWMAC_IPQ806X && DWMAC_LPC18XX && DWMAC_ROCKCHIP && DWMAC_SOCFPGA && DWMAC_STI && TI_CPSW && PINCTRL_ROCKCHIP && PINCTRL_DOVE && POWER_RESET_KEYSTONE && S3C2410_WATCHDOG && VIDEO_OMAP3 && VIDEO_S5P_FIMC && USB_XHCI_MTK && RTC_DRV_AT91SAM9 && LPC18XX_DMAMUX && VIDEO_OMAP4 && HWSPINLOCK_QCOM && ATMEL_ST && QCOM_GSBI && PHY_HI6220_USB) selects MFD_SYSCON which has unmet direct dependencies (HAS_IOMEM) > > Reuse the !MMU variants for !HAS_IOMEM as they are sufficient for our > needs. This fixes build errors for UM allyesconfig: > > drivers/mfd/syscon.c:89:2: error: implicit declaration of function 'iounmap' [-Werror=implicit-function-declaration] > iounmap(base); > > Reported-by: Fengguang Wu <fengguang.wu@intel.com> > Cc: Arnd Bergmann <arnd@arndb.de> > Cc: linux-arch@vger.kernel.org > Signed-off-by: Rob Herring <robh@kernel.org> > --- > include/asm-generic/io.h | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) Some time ago I posted patches adding the dependency, but that path looks like a better approach: Tested-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Best regards, Krzysztof
Am 29.03.2016 um 22:13 schrieb Rob Herring: >>> Reuse the !MMU variants for !HAS_IOMEM as they are sufficient for our >>> needs. This fixes build errors for UM allyesconfig: >>> >>> drivers/mfd/syscon.c:89:2: error: implicit declaration of function 'iounmap' [-Werror=implicit-function-declaration] >>> iounmap(base); >>> >>> Reported-by: Fengguang Wu <fengguang.wu@intel.com> >>> Cc: Arnd Bergmann <arnd@arndb.de> >>> Cc: linux-arch@vger.kernel.org >>> Signed-off-by: Rob Herring <robh@kernel.org> >> >> (adding Richard and the UML list to cc) >> >> I actually prototyped a patch that did the opposite: remove the readl/writel/... >> definitions when HAS_IOMEM is unset. I didn't get far enough to submit it, >> but see below for what I did. > > Ewww. Why do the opposite of what we do for every other Kconfig symbol > which is provide empty functions? It really only functions as a > disable on UML flag which runs counter to enabling drivers to build > for all arches. > > I actually started a patch to remove the HAS_IOMEM dependency > everywhere (or just the per driver cases). It didn't break as bad as I > expected, but became more than I wanted to fix. Mainly, all the devm_ > variants also need empty versions or to be always enabled. The root cause of the problem is COMPILE_TEST. People who use it need to get forced to think about the consequences. COMPILE_TEST means that the driver will also be build on not so fancy archs like UML, s390 and m68k where you don't have IOMEM or also DMA in every possible configuration. So a quick build test on x86 is not sufficient. This is why I'm absolutely not a fan of having stubs for iounmap() an friends for UML. It only tries to bypass the root cause. What is next? Stubs for DMA? PCI? For everything else that does not build? With COMPILE_TEST we have created a monster and now we have to deal with it in terms of having correct dependencies when COMPILE_TEST is being used. Thanks, //richard
On Wednesday 30 March 2016 09:50:22 Richard Weinberger wrote: > Am 29.03.2016 um 22:13 schrieb Rob Herring: > >>> Reuse the !MMU variants for !HAS_IOMEM as they are sufficient for our > >>> needs. This fixes build errors for UM allyesconfig: > >>> > >>> drivers/mfd/syscon.c:89:2: error: implicit declaration of function 'iounmap' [-Werror=implicit-function-declaration] > >>> iounmap(base); > >>> > >>> Reported-by: Fengguang Wu <fengguang.wu@intel.com> > >>> Cc: Arnd Bergmann <arnd@arndb.de> > >>> Cc: linux-arch@vger.kernel.org > >>> Signed-off-by: Rob Herring <robh@kernel.org> > >> > >> (adding Richard and the UML list to cc) > >> > >> I actually prototyped a patch that did the opposite: remove the readl/writel/... > >> definitions when HAS_IOMEM is unset. I didn't get far enough to submit it, > >> but see below for what I did. > > > > Ewww. Why do the opposite of what we do for every other Kconfig symbol > > which is provide empty functions? It really only functions as a > > disable on UML flag which runs counter to enabling drivers to build > > for all arches. > > > > I actually started a patch to remove the HAS_IOMEM dependency > > everywhere (or just the per driver cases). It didn't break as bad as I > > expected, but became more than I wanted to fix. Mainly, all the devm_ > > variants also need empty versions or to be always enabled. > > The root cause of the problem is COMPILE_TEST. People who use it need to > get forced to think about the consequences. > COMPILE_TEST means that the driver will also be build on not so fancy archs > like UML, s390 and m68k where you don't have IOMEM or also DMA in every > possible configuration. > So a quick build test on x86 is not sufficient. > > This is why I'm absolutely not a fan of having stubs for iounmap() an friends > for UML. It only tries to bypass the root cause. > What is next? Stubs for DMA? PCI? For everything else that does not build? > > With COMPILE_TEST we have created a monster and now we have to deal with it in terms of > having correct dependencies when COMPILE_TEST is being used. > One way out of this would be accept that UML is just too different from the other architectures, and make COMPILE_TEST itself depend on !UML. COMPILE_TEST is highly valuable for me, because it means that all the ARM specific drivers can now be build-tested on x86 and they show up for anyone doing allmodconfig builds. If we introduce a regression in any of the drivers, it gets caught much quicker because many subsystem maintainers run an x86 allmodconfig build after pulling downstream maintainer trees. There is also stuff like Coverity that only ever runs on x86. It does make sense to have COMPILE_TEST imply that we are able to build things on all architectures, not just x86 in addition to the platform a driver is meant for, but the value of compile-testing a random driver with UML in particular approaches zero, and instead causes lots of work. Arnd
Am 30.03.2016 um 10:04 schrieb Arnd Bergmann: > On Wednesday 30 March 2016 09:50:22 Richard Weinberger wrote: >> Am 29.03.2016 um 22:13 schrieb Rob Herring: >>>>> Reuse the !MMU variants for !HAS_IOMEM as they are sufficient for our >>>>> needs. This fixes build errors for UM allyesconfig: >>>>> >>>>> drivers/mfd/syscon.c:89:2: error: implicit declaration of function 'iounmap' [-Werror=implicit-function-declaration] >>>>> iounmap(base); >>>>> >>>>> Reported-by: Fengguang Wu <fengguang.wu@intel.com> >>>>> Cc: Arnd Bergmann <arnd@arndb.de> >>>>> Cc: linux-arch@vger.kernel.org >>>>> Signed-off-by: Rob Herring <robh@kernel.org> >>>> >>>> (adding Richard and the UML list to cc) >>>> >>>> I actually prototyped a patch that did the opposite: remove the readl/writel/... >>>> definitions when HAS_IOMEM is unset. I didn't get far enough to submit it, >>>> but see below for what I did. >>> >>> Ewww. Why do the opposite of what we do for every other Kconfig symbol >>> which is provide empty functions? It really only functions as a >>> disable on UML flag which runs counter to enabling drivers to build >>> for all arches. >>> >>> I actually started a patch to remove the HAS_IOMEM dependency >>> everywhere (or just the per driver cases). It didn't break as bad as I >>> expected, but became more than I wanted to fix. Mainly, all the devm_ >>> variants also need empty versions or to be always enabled. >> >> The root cause of the problem is COMPILE_TEST. People who use it need to >> get forced to think about the consequences. >> COMPILE_TEST means that the driver will also be build on not so fancy archs >> like UML, s390 and m68k where you don't have IOMEM or also DMA in every >> possible configuration. >> So a quick build test on x86 is not sufficient. >> >> This is why I'm absolutely not a fan of having stubs for iounmap() an friends >> for UML. It only tries to bypass the root cause. >> What is next? Stubs for DMA? PCI? For everything else that does not build? >> >> With COMPILE_TEST we have created a monster and now we have to deal with it in terms of >> having correct dependencies when COMPILE_TEST is being used. >> > > One way out of this would be accept that UML is just too different > from the other architectures, and make COMPILE_TEST itself depend > on !UML. > > COMPILE_TEST is highly valuable for me, because it means that all the > ARM specific drivers can now be build-tested on x86 and they show up > for anyone doing allmodconfig builds. If we introduce a regression > in any of the drivers, it gets caught much quicker because many > subsystem maintainers run an x86 allmodconfig build after pulling > downstream maintainer trees. There is also stuff like Coverity that > only ever runs on x86. > > It does make sense to have COMPILE_TEST imply that we are able > to build things on all architectures, not just x86 in addition to > the platform a driver is meant for, but the value of compile-testing > a random driver with UML in particular approaches zero, and instead > causes lots of work. I fully understand your point of view. COMPILE_TEST is a monster that can do the heavy lifting for you, but monsters also have claws and fangs. ;-) Having COMPILE_TEST having depend on !UML works for me. But don't we have other archs without io mem? At least a few years ago while porting nandsim to UML I found s390 that lacks of io mem too. Maybe we depend COMPILE_TEST on HAS_IOMEM? Thanks, //richard
On Wednesday 30 March 2016 10:13:53 Richard Weinberger wrote: > > I fully understand your point of view. COMPILE_TEST is a monster that > can do the heavy lifting for you, but monsters also have claws and fangs. > > Having COMPILE_TEST having depend on !UML works for me. But don't > we have other archs without io mem? At least a few years ago while > porting nandsim to UML I found s390 that lacks of io mem too. s390 gained IOMEM support when they started having PCI attachments. arch/score selects NO_IOMEM, though they do in fact use MMIO, and I'm sure their architecture has lots of other problems with build testing that nobody cares about. arch/tile can select NO_IOMEM when PCI is disabled, they might care about this, though I think they also have other build-time issues. > Maybe we depend COMPILE_TEST on HAS_IOMEM? That sounds fine with me as well. Arnd
On Wed, Mar 30, 2016 at 5:03 AM, Arnd Bergmann <arnd@arndb.de> wrote: > On Wednesday 30 March 2016 10:13:53 Richard Weinberger wrote: >> >> I fully understand your point of view. COMPILE_TEST is a monster that >> can do the heavy lifting for you, but monsters also have claws and fangs. >> >> Having COMPILE_TEST having depend on !UML works for me. But don't >> we have other archs without io mem? At least a few years ago while >> porting nandsim to UML I found s390 that lacks of io mem too. > > s390 gained IOMEM support when they started having PCI attachments. I'm confused how s390 defines HAS_IOMEM and includes it from lib/Kconfig though. > arch/score selects NO_IOMEM, though they do in fact use MMIO, and > I'm sure their architecture has lots of other problems with build > testing that nobody cares about. > > arch/tile can select NO_IOMEM when PCI is disabled, they might > care about this, though I think they also have other build-time > issues. > >> Maybe we depend COMPILE_TEST on HAS_IOMEM? > > That sounds fine with me as well. That or !UML is fine, but I still think we should rid drivers from depending on HAS_IOMEM. We probably still have cases where the only dependency besides the driver's subsystem is HAS_IOMEM. Rob
On Wednesday 30 March 2016 08:29:45 Rob Herring wrote: > On Wed, Mar 30, 2016 at 5:03 AM, Arnd Bergmann <arnd@arndb.de> wrote: > > On Wednesday 30 March 2016 10:13:53 Richard Weinberger wrote: > >> > >> I fully understand your point of view. COMPILE_TEST is a monster that > >> can do the heavy lifting for you, but monsters also have claws and fangs. > >> > >> Having COMPILE_TEST having depend on !UML works for me. But don't > >> we have other archs without io mem? At least a few years ago while > >> porting nandsim to UML I found s390 that lacks of io mem too. > > > > s390 gained IOMEM support when they started having PCI attachments. > > I'm confused how s390 defines HAS_IOMEM and includes it from lib/Kconfig though. Kconfig can define the same symbol multiple times, which is really confusing. I only checked for NO_IOMEM, which s390 no longer defines, but I have not checked what actually happens in case of two conflicting definitions, where the one lib/Kconfig defaults to 'y' and the one in arch/s390 defaults to CONFIG_PCI. Arnd
Hi Rob, On Tue, Mar 29, 2016 at 10:13 PM, Rob Herring <robh@kernel.org> wrote: > On Tue, Mar 29, 2016 at 2:37 PM, Arnd Bergmann <arnd@arndb.de> wrote: >> On Tuesday 29 March 2016 13:23:00 Rob Herring wrote: >>> Drivers shouldn't have to care about HAS_IOMEM to compile and having to >>> causes a Kconfig mess: >>> >>> warning: (MEDIA_SUBDRV_AUTOSELECT && VIDEO_CX231XX && INV_MPU6050_I2C) selects I2C_MUX which has unmet direct dependencies (I2C && HAS_IOMEM) >>> warning: (ST_IRQCHIP && STMMAC_PLATFORM && DWMAC_IPQ806X && DWMAC_LPC18XX && DWMAC_ROCKCHIP && DWMAC_SOCFPGA && DWMAC_STI && TI_CPSW && PINCTRL_ROCKCHIP && PINCTRL_DOVE && POWER_RESET_KEYSTONE && S3C2410_WATCHDOG && VIDEO_OMAP3 && VIDEO_S5P_FIMC && USB_XHCI_MTK && RTC_DRV_AT91SAM9 && LPC18XX_DMAMUX && VIDEO_OMAP4 && HWSPINLOCK_QCOM && ATMEL_ST && QCOM_GSBI && PHY_HI6220_USB) selects MFD_SYSCON which has unmet direct dependencies (HAS_IOMEM) >>> >>> Reuse the !MMU variants for !HAS_IOMEM as they are sufficient for our >>> needs. This fixes build errors for UM allyesconfig: >>> >>> drivers/mfd/syscon.c:89:2: error: implicit declaration of function 'iounmap' [-Werror=implicit-function-declaration] >>> iounmap(base); >>> >>> Reported-by: Fengguang Wu <fengguang.wu@intel.com> >>> Cc: Arnd Bergmann <arnd@arndb.de> >>> Cc: linux-arch@vger.kernel.org >>> Signed-off-by: Rob Herring <robh@kernel.org> >> >> (adding Richard and the UML list to cc) >> >> I actually prototyped a patch that did the opposite: remove the readl/writel/... >> definitions when HAS_IOMEM is unset. I didn't get far enough to submit it, >> but see below for what I did. > > Ewww. Why do the opposite of what we do for every other Kconfig symbol > which is provide empty functions? It really only functions as a > disable on UML flag which runs counter to enabling drivers to build > for all arches. Usually the empty function fall into one of two classes: 1. They return an error, so drivers will abort their initialization, 2. They are a plain no-op, for functions with harmless side-effects. The !MMU versions are not dummies, but assume a transparent translation. This may lead to drivers continuing their initialization and crashing the system. > I actually started a patch to remove the HAS_IOMEM dependency > everywhere (or just the per driver cases). It didn't break as bad as I > expected, but became more than I wanted to fix. Mainly, all the devm_ > variants also need empty versions or to be always enabled. Should these dependencies on HAS_IOMEM be changed to "HAS_IOMEM || COMPILE_TEST"? Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds
Am 30.03.2016 um 22:08 schrieb Geert Uytterhoeven: >> I actually started a patch to remove the HAS_IOMEM dependency >> everywhere (or just the per driver cases). It didn't break as bad as I >> expected, but became more than I wanted to fix. Mainly, all the devm_ >> variants also need empty versions or to be always enabled. > > Should these dependencies on HAS_IOMEM be changed to "HAS_IOMEM || > COMPILE_TEST"? Another idea, the issue with COMPILE_TEST is that it is automatically selected by make targets such as allyesconfig or allmodconfig. Couldn't we make COMPILE_TEST just harder to select? i.e. that with allyesconfig it is still disabled and if someone really wants to do a compile test of anything we could issue something like "make allyesconfig COMPILE_TEST=y" to have COMPILE_TEST selected? Thanks, //richard
On Wednesday 30 March 2016 22:35:37 Richard Weinberger wrote: > Am 30.03.2016 um 22:08 schrieb Geert Uytterhoeven: > >> I actually started a patch to remove the HAS_IOMEM dependency > >> everywhere (or just the per driver cases). It didn't break as bad as I > >> expected, but became more than I wanted to fix. Mainly, all the devm_ > >> variants also need empty versions or to be always enabled. > > > > Should these dependencies on HAS_IOMEM be changed to "HAS_IOMEM || > > COMPILE_TEST"? > > Another idea, the issue with COMPILE_TEST is that it is automatically selected > by make targets such as allyesconfig or allmodconfig. > Couldn't we make COMPILE_TEST just harder to select? > i.e. that with allyesconfig it is still disabled and if someone really wants to > do a compile test of anything we could issue something like "make allyesconfig COMPILE_TEST=y" > to have COMPILE_TEST selected? > I think it's essential for allmodconfig and allyesconfig to enable COMPILE_TEST: the only reason for ever building those configurations is for compile tests. Arnd
On Wed, Mar 30, 2016 at 3:08 PM, Geert Uytterhoeven <geert@linux-m68k.org> wrote: > Hi Rob, > > On Tue, Mar 29, 2016 at 10:13 PM, Rob Herring <robh@kernel.org> wrote: >> On Tue, Mar 29, 2016 at 2:37 PM, Arnd Bergmann <arnd@arndb.de> wrote: >>> On Tuesday 29 March 2016 13:23:00 Rob Herring wrote: >>>> Drivers shouldn't have to care about HAS_IOMEM to compile and having to >>>> causes a Kconfig mess: >>>> >>>> warning: (MEDIA_SUBDRV_AUTOSELECT && VIDEO_CX231XX && INV_MPU6050_I2C) selects I2C_MUX which has unmet direct dependencies (I2C && HAS_IOMEM) >>>> warning: (ST_IRQCHIP && STMMAC_PLATFORM && DWMAC_IPQ806X && DWMAC_LPC18XX && DWMAC_ROCKCHIP && DWMAC_SOCFPGA && DWMAC_STI && TI_CPSW && PINCTRL_ROCKCHIP && PINCTRL_DOVE && POWER_RESET_KEYSTONE && S3C2410_WATCHDOG && VIDEO_OMAP3 && VIDEO_S5P_FIMC && USB_XHCI_MTK && RTC_DRV_AT91SAM9 && LPC18XX_DMAMUX && VIDEO_OMAP4 && HWSPINLOCK_QCOM && ATMEL_ST && QCOM_GSBI && PHY_HI6220_USB) selects MFD_SYSCON which has unmet direct dependencies (HAS_IOMEM) >>>> >>>> Reuse the !MMU variants for !HAS_IOMEM as they are sufficient for our >>>> needs. This fixes build errors for UM allyesconfig: >>>> >>>> drivers/mfd/syscon.c:89:2: error: implicit declaration of function 'iounmap' [-Werror=implicit-function-declaration] >>>> iounmap(base); >>>> >>>> Reported-by: Fengguang Wu <fengguang.wu@intel.com> >>>> Cc: Arnd Bergmann <arnd@arndb.de> >>>> Cc: linux-arch@vger.kernel.org >>>> Signed-off-by: Rob Herring <robh@kernel.org> >>> >>> (adding Richard and the UML list to cc) >>> >>> I actually prototyped a patch that did the opposite: remove the readl/writel/... >>> definitions when HAS_IOMEM is unset. I didn't get far enough to submit it, >>> but see below for what I did. >> >> Ewww. Why do the opposite of what we do for every other Kconfig symbol >> which is provide empty functions? It really only functions as a >> disable on UML flag which runs counter to enabling drivers to build >> for all arches. > > Usually the empty function fall into one of two classes: > 1. They return an error, so drivers will abort their initialization, > 2. They are a plain no-op, for functions with harmless side-effects. > > The !MMU versions are not dummies, but assume a transparent translation. > This may lead to drivers continuing their initialization and crashing the > system. I thought about this, but how could you even get to the point of having some physical address passed to the driver? Perhaps some old ISA driver, but I'm not sure memory ranges were ever fixed (only i/o). We're really only talking about UML here. It can be done though, it's just more ifdef'ery. >> I actually started a patch to remove the HAS_IOMEM dependency >> everywhere (or just the per driver cases). It didn't break as bad as I >> expected, but became more than I wanted to fix. Mainly, all the devm_ >> variants also need empty versions or to be always enabled. > > Should these dependencies on HAS_IOMEM be changed to "HAS_IOMEM || > COMPILE_TEST"? Why? Most are either HAS_IOMEM or HAS_IOMEM && COMPILE_TEST. Rob
Hi Rob, On Wed, Mar 30, 2016 at 11:20 PM, Rob Herring <robh@kernel.org> wrote: > On Wed, Mar 30, 2016 at 3:08 PM, Geert Uytterhoeven > <geert@linux-m68k.org> wrote: >> On Tue, Mar 29, 2016 at 10:13 PM, Rob Herring <robh@kernel.org> wrote: >>> Ewww. Why do the opposite of what we do for every other Kconfig symbol >>> which is provide empty functions? It really only functions as a >>> disable on UML flag which runs counter to enabling drivers to build >>> for all arches. >> >> Usually the empty function fall into one of two classes: >> 1. They return an error, so drivers will abort their initialization, >> 2. They are a plain no-op, for functions with harmless side-effects. >> >> The !MMU versions are not dummies, but assume a transparent translation. >> This may lead to drivers continuing their initialization and crashing the >> system. > > I thought about this, but how could you even get to the point of > having some physical address passed to the driver? Perhaps some old > ISA driver, but I'm not sure memory ranges were ever fixed (only i/o). > We're really only talking about UML here. Hmm, perhaps it indeed can't happen. >>> I actually started a patch to remove the HAS_IOMEM dependency >>> everywhere (or just the per driver cases). It didn't break as bad as I >>> expected, but became more than I wanted to fix. Mainly, all the devm_ >>> variants also need empty versions or to be always enabled. >> >> Should these dependencies on HAS_IOMEM be changed to "HAS_IOMEM || >> COMPILE_TEST"? > > Why? Most are either HAS_IOMEM or HAS_IOMEM && COMPILE_TEST. COMPILE_TEST is meant to enable drivers that are not useful, due to missing infrastructure. Without having it enabled, you don't want to include useless drivers. If you have dummies, you still want the HAS_IOMEM dependency to avoid compiling useless drivers. I.e. it's a hint for distros, who don't want to ship useless drivers. Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds
diff --git a/include/asm-generic/io.h b/include/asm-generic/io.h index eed3bbe..03160b0 100644 --- a/include/asm-generic/io.h +++ b/include/asm-generic/io.h @@ -753,7 +753,7 @@ static inline void *phys_to_virt(unsigned long address) * ioremap_*() variant as defined to itself to avoid the default NULL return. */ -#ifdef CONFIG_MMU +#if defined(CONFIG_MMU) && defined(CONFIG_HAS_IOMEM) #ifndef ioremap_uc #define ioremap_uc ioremap_uc
Drivers shouldn't have to care about HAS_IOMEM to compile and having to causes a Kconfig mess: warning: (MEDIA_SUBDRV_AUTOSELECT && VIDEO_CX231XX && INV_MPU6050_I2C) selects I2C_MUX which has unmet direct dependencies (I2C && HAS_IOMEM) warning: (ST_IRQCHIP && STMMAC_PLATFORM && DWMAC_IPQ806X && DWMAC_LPC18XX && DWMAC_ROCKCHIP && DWMAC_SOCFPGA && DWMAC_STI && TI_CPSW && PINCTRL_ROCKCHIP && PINCTRL_DOVE && POWER_RESET_KEYSTONE && S3C2410_WATCHDOG && VIDEO_OMAP3 && VIDEO_S5P_FIMC && USB_XHCI_MTK && RTC_DRV_AT91SAM9 && LPC18XX_DMAMUX && VIDEO_OMAP4 && HWSPINLOCK_QCOM && ATMEL_ST && QCOM_GSBI && PHY_HI6220_USB) selects MFD_SYSCON which has unmet direct dependencies (HAS_IOMEM) Reuse the !MMU variants for !HAS_IOMEM as they are sufficient for our needs. This fixes build errors for UM allyesconfig: drivers/mfd/syscon.c:89:2: error: implicit declaration of function 'iounmap' [-Werror=implicit-function-declaration] iounmap(base); Reported-by: Fengguang Wu <fengguang.wu@intel.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: linux-arch@vger.kernel.org Signed-off-by: Rob Herring <robh@kernel.org> --- include/asm-generic/io.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) -- 2.5.0