[OpenWrt-Devel] Cleanup in 4.19 generic kernel config?

[Daniel Engberg]

On 2019-02-13 00:08, Rosen Penev wrote:
> On Sat, Feb 9, 2019 at 12:16 AM Daniel Engberg
> <daniel.engberg.lists at pyret.net> wrote:
>> 
>> Hi,
>> 
>> After my attempt getting Linux 4.19 working for the Octeon target
>> I started to look at the generic configuration for 4.19 and it
>> seems that there are quite a bit of target specific options listed
>> and potentially generating larger kernels than necessary.
>> Given that some options are incorrect for certain targets I assume
>> that they get overridden and/or ignored. Unless I misunderstand
>> how the Linux kernel configuration works =y is equivalent to always
>> compiled into kernel?
>> 
>> I'm taking about this file by the way:
>> https://github.com/openwrt/openwrt/blob/master/target/linux/generic/config-4.19
>> 
>> Should be set by target? Shouldn't coexist with CONFIG_64BIT as
>> far as I can tell.
>> CONFIG_32BIT=y
>> 
>> ARM specific, should be set by target?
>> CONFIG_ARM_CPU_TOPOLOGY=y
>> CONFIG_ARM_DMA_MEM_BUFFERABLE=y
>> CONFIG_ARM_GIC_MAX_NR=1
>> 
>> Target specific, no need to have it enabled by default and/or
>> possibly at all?
>> CONFIG_ATA_BMDMA=y
>> CONFIG_ATA_SFF=y
>> 
>> Seems to be defined by default?
>> CONFIG_BLK_DEV_LOOP_MIN_COUNT=8
>> https://github.com/torvalds/linux/blob/master/drivers/block/Kconfig#L217
>> 
>> Should be set by target?
>> CONFIG_BOOKE_WDT_DEFAULT_TIMEOUT=3
>> 
>> Debug option - Adds unnecessary overhead?
>> CONFIG_BRANCH_PROFILE_NONE=y
>> https://cateee.net/lkddb/web-lkddb/BRANCH_PROFILE_NONE.html
>> 
>> Should be set by target (ARM)?
>> CONFIG_CACHE_L2X0_PMU=y
>> 
>> Should be set by target, why is it enabled in 2019?
>> CONFIG_CARDBUS=y
>> 
>> Should be set by target (seems x86/64x specific)?
>> CONFIG_DOUBLEFAULT=y
>> 
>> Gets defined automatically based on target/platform?
>> CONFIG_HZ=100
>> CONFIG_HZ_100=y
>> 
>> Already default?
>> CONFIG_IIO_CONSUMERS_PER_TRIGGER=2
>> https://cateee.net/lkddb/web-lkddb/IIO_CONSUMERS_PER_TRIGGER.html
>> 
>> Why?
>> CONFIG_ISDN=y
>> CONFIG_JOLIET=y
>> 
>> Should be set by target (ARM)?
>> CONFIG_KERNEL_MODE_NEON=y
>> 
>> Set Page_Size by target?
>> CONFIG_PAGE_SIZE_4KB=y
>> CONFIG_PPC_4K_PAGES=y
>> 
>> Already default?
>> CONFIG_PWRSEQ_EMMC=y
>> https://cateee.net/lkddb/web-lkddb/PWRSEQ_EMMC.html
>> 
>> Default 21 isn't enough?
>> CONFIG_RCU_CPU_STALL_TIMEOUT=60
>> 
>> Needed for?
>> CONFIG_RXKAD=y
>> 
>> Obsolete?
>> CONFIG_SLABINFO=y
>> 
>> Shouldn't be default and/or platform specific?
>> CONFIG_SND_X86=y
>> 
>> Needed for?
>> CONFIG_USB_ARMLINUX=y
>> CONFIG_USB_BELKIN=y
>> 
>> Platform specific?
>> CONFIG_USB_OHCI_LITTLE_ENDIAN=y
>> 
>> Platform specific?
>> CONFIG_X86_SYSFB=y
>> 
>> Best regads,
>> Daniel
> I've been running this with ramips for a few days. No issues to report.
>> 
>> _______________________________________________
>> openwrt-devel mailing list
>> openwrt-devel at lists.openwrt.org
>> https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Hi,

Thanks for giving it a try, hopefully someone we can make generic a
little bit cleaner in the end. Since you tried it on ramips and going
by this 
https://github.com/neheb/source/commit/500f88b54beb9399d1e4927fcd2d0fc59154dfe1
I'm going to assume that you tried this on 4.14? :-)

Best regards,
Daniel

_______________________________________________
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel