Bug 216748
| Summary: | Kconfig has some dangling references | ||
|---|---|---|---|
| Product: | Other | Reporter: | Andrew Jones (andrew.jones) |
| Component: | Configuration | Assignee: | other_configuration (other_configuration) |
| Status: | NEW --- | ||
| Severity: | normal | CC: | andy.shevchenko, arnd |
| Priority: | P1 | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Kernel Version: | Subsystem: | ||
| Regression: | No | Bisected commit-id: | |
| Attachments: |
perl script to hunt for dangling references in Kconfig
grep results for each symbol output from the attached perl script |
||
Created attachment 303312 [details]
grep results for each symbol output from the attached perl script
for s in $(./bad-syms.pl); do echo $s ---; git grep -n -w -e $s -e CONFIG_$s; echo; done
PL310 had a bunch of irrelevant lines in the grep output that I manually removed.
I believe the best is to drop a message into Linux Kernel Mailing List for that. And actually discuss there that somebody will take care of bugzilla.kernel.org for risc-V issues. +Cc: Arnd for the best of this effort. |
Created attachment 303311 [details] perl script to hunt for dangling references in Kconfig A recent patch for arch/riscv/Kconfig pointed out that the wrong name for a symbol was being referenced in 'range' attributes. This implied that other "dangling references" may exist. I'm not aware of a kconfig check which can be run for this detection so I hacked an ugly perl script to attempt to do it. Here are the results of running it ARCH_BCM4908 ARCH_HAS_HOLES_MEMORYMODEL ARCH_THUNDERBAY ARCH_VERSAL ARM_ERRATA_794072 CC_CAN_LINK_STATIC_NO_RUNTIME_DEPS CLK_FIXED_FCH CONFIG_STM DEBUG_OMAP3UART1 DEBUG_OMAP3UART2 DRM_DEBUG_SELFTEST DRM_DP_HELPER DRM_KMS_DMA_HELPER HAVE_LEGACY_CLK INTEL_IFS_DEVICE MACH_DAVINCI_DM6467_EVM MACH_DAVINCI_EVM MACH_JZ4760 MFD_MAX597X MFD_SM5703 MFD_TN48M_CPLD MFD_TPS65219 MIPS_BAIKAL_T1 MTK_DVFSRC PINCTRL_MILBEAUT PL310 SBI_V01 SND_SOC_AC97_BUS_NEW USB_HSIC_USB3613 SBI_V01 was the one pointed out in the recent riscv patch. And checking with grep $ git grep -n -w -e SBI_V01 -e CONFIG_SBI_V01 arch/riscv/Kconfig:320: range 2 512 if !SBI_V01 arch/riscv/Kconfig:321: range 2 32 if SBI_V01 && 32BIT arch/riscv/Kconfig:322: range 2 64 if SBI_V01 && 64BIT shows the symbol is indeed non-exist and never used. Running the same grep on all the symbols produced output that appears to validate the perl script (which is attached). I'm not exactly sure how to alert all the different drivers and subsystems about potential issues with their respective symbols. Below is a list of affected files and I'll attach the grep output for each symbol like the SBI_V01 example above. Hopefully the filenames will trigger some filters. (I've removed the riscv output since it's already known and fixed.) arch/arm/Kconfig.debug arch/arm/mach-milbeaut/Kconfig arch/arm/mach-npcm/Kconfig arch/arm/mach-omap1/Kconfig arch/arm/mach-socfpga/Kconfig arch/arm/tools/mach-types arch/m68k/Kconfig.cpu arch/mips/Kconfig arch/sh/boards/Kconfig arch/um/Kconfig drivers/ata/Kconfig drivers/bus/Kconfig drivers/clk/baikal-t1/Kconfig drivers/clk/ingenic/Kconfig drivers/clk/Kconfig drivers/gpio/Kconfig drivers/gpu/drm/bridge/Kconfig drivers/gpu/drm/i915/Kconfig.debug drivers/gpu/drm/logicvc/Kconfig drivers/hwmon/Kconfig drivers/hwtracing/stm/Kconfig drivers/hwtracing/stm/Makefile drivers/i2c/busses/Kconfig drivers/leds/blink/Kconfig drivers/Makefile drivers/memory/Kconfig drivers/mtd/maps/Kconfig drivers/mtd/parsers/Kconfig drivers/phy/intel/Kconfig drivers/pinctrl/Kconfig drivers/pinctrl/pinctrl-ingenic.c drivers/platform/x86/intel/ifs/Kconfig drivers/regulator/Kconfig drivers/reset/Kconfig drivers/spi/Kconfig drivers/staging/greybus/arche-platform.c drivers/staging/greybus/Kconfig drivers/usb/dwc3/Kconfig sound/soc/amd/acp/Kconfig sound/soc/amd/Kconfig sound/soc/pxa/Kconfig sound/soc/ti/Kconfig