generic: track and understand wrong formatted patch in hack/pending

Ansuel commented 2 years ago

Hi,

This is not a real issue with Openwrt itself but more of a problem we have on our repo.

We have too many patches that are wrongly formatted, with no author/sob, with no reason they are actually there and sometimes (probably) even problematic with recent changes to kernel version.

We have to fix this for good or at least have a place to track them and with times try to fix them.

So for every maintainer, feel free to edit this and add additional info about patch status...

(this is a wip issue so i'm adding stuff as i find more info)

THE LIST

hack-5.15

204-module_strip.patch (OK. an actualy hack... in theory this should be reworked to kernel symbol whitelisting...)
205-kconfig-exit.patch (https://www.mail-archive.com/openwrt-devel@lists.openwrt.org/msg59501.html)
210-darwin_scripts_include.patch
211-darwin-uuid-typedef-clash.patch
212-tools_portability.patch
214-spidev_h_portability.patch
220-arm-gc_sections.patch (to strip kernel modules)
221-module_exports.patch (from @nbd168 its purpose is to allow the build system to generate a list of needed module exports and remove any unnecessary exports from the kernel image)
230-openwrt_lzma_options.patch
249-udp-tunnel-selection.patch (can be dropped the udp_tunnel ko is compiled on selecting vxlan, on openwrt is used only by that package)
250-netfilter_depends.patch
251-kconfig.patch
252-SATA_PMP.patch
253-ksmbd-config.patch
259-regmap_dynamic.patch (can be upstreamed but i need authorization from @nbd168)
260-crypto_test_dependencies.patch (half of it is already present upstream... only kconfig changes are required... need @nbd168 authorization to propose them)
261-lib-arc4-unhide.patch (needed for backports... wireless package require arc4 libs for WEP)
280-rfkill-stubs.patch (can be dropped, new kernel already handle this with another implementation (depends RFKILL || !RFKILL)
300-MIPS-r4k_cache-use-more-efficient-cache-blast.patch
301-mips_image_cmdline_hack.patch
321-powerpc_crtsavres_prereq.patch
402-mtd-blktrans-call-add-disks-after-mtd-device.patch
410-block-fit-partition-parser.patch
420-mtd-set-rootfs-to-be-root-dev.patch
421-drivers-mtd-parsers-add-nvmem-support-to-cmdlinepart.patch (to drop we have a solution upstream but major rework are needed aka WIP https://github.com/openwrt/openwrt/pull/10153)
430-mtk-bmt-support.patch
531-debloat_lzma.patch
600-bridge_offload.patch
645-netfilter-connmark-introduce-set-dscpmark.patch
650-netfilter-add-xt_FLOWOFFLOAD-target.patch
651-wireless_mesh_header.patch (required for backports project... replace it with a flag to force the value?)
660-fq_codel_defaults.patch (drop this? openwrt still supports 32mb ram devices?)
661-kernel-ct-size-the-hashtable-more-adequately.patch
700-swconfig_switch_drivers.patch
711-net-dsa-mv88e6xxx-disable-ATU-violation.patch (worth investigate this...)
720-net-phy-add-aqr-phys.patch (why never sent upstream?)
721-net-add-packet-mangeling.patch
722-net-phy-aquantia-enable-AQR112-and-AQR412.patch (why never sent upstream?)
723-net-phy-aquantia-fix-system-side-protocol-mi.patch (why never sent upstream?)
724-net-phy-aquantia-Add-AQR113-driver-support.patch (why never sent upstream?)
725-net-phy-aquantia-add-PHY_IDs-for-AQR112-variants.patch (why never sent upstream?)
760-net-usb-r8152-add-LED-configuration-from-OF.patch (won't ever happen... upstream require a totally different implementation)
761-dt-bindings-net-add-RTL8152-binding-documentation.patch (@blocktrron can you rework this and try to send at least part of the Documentation? aka the patch without the led part?)
773-bgmac-add-srab-switch.patch (@hauke chance this can be reworked and sent upstream? by getting the reg from a compatible node instead of hardcoding the ioremap?)
780-usb-net-MeigLink_modem_support.patch (why never sent upstream?)
800-GPIO-add-named-gpio-exports.patch
901-debloat_sock_diag.patch
902-debloat_proc.patch
904-debloat_dma_buf.patch
910-kobject_uevent.patch
911-kobject_add_broadcast_uevent.patch
920-device_tree_cmdline.patch

Ansuel commented 2 years ago

Ok... I added all the comments for the patch that i think can be improved... hoping we will get some answer...

blocktrron commented 2 years ago

761-dt-bindings-net-add-RTL8152-binding-documentation.patch

There is no need for the Documentation without the OF part. We might as well just drop this patch altogether, as it isn't going upstream.

Ansuel commented 2 years ago

There is no need for the Documentation without the OF part. We might as well just drop this patch altogether, as it isn't going upstream.

The idea would be to push a simple Documentation and then if we are lucky adding the leds relevant part when it will be ready.

blocktrron commented 2 years ago

There is no need for documentation without the LED binding. I have no interest in investing time to push a stub upstream.

hauke commented 2 years ago

I think this patch is not needed any more: https://github.com/openwrt/openwrt/blob/master/target/linux/generic/hack-5.15/773-bgmac-add-srab-switch.patch

The b53 swconfig srab is not used any more, bcm53xx switched to the DSA driver.

$ git grep SWCONFIG_B53_SRAB_DRIVER
target/linux/bcm47xx/config-5.10:# CONFIG_SWCONFIG_B53_SRAB_DRIVER is not set
target/linux/generic/config-5.10:# CONFIG_SWCONFIG_B53_SRAB_DRIVER is not set
target/linux/generic/config-5.15:# CONFIG_SWCONFIG_B53_SRAB_DRIVER is not set
target/linux/generic/files/drivers/net/phy/b53/Kconfig:config SWCONFIG_B53_SRAB_DRIVER
target/linux/generic/files/drivers/net/phy/b53/Makefile:obj-$(CONFIG_SWCONFIG_B53_SRAB_DRIVER)  += b53_srab.o
target/linux/mpc85xx/p1020/config-default:# CONFIG_SWCONFIG_B53_SRAB_DRIVER is not set

neheb commented 2 years ago

r4k cache blast patch comes from an old QCA SDK.

openwrt / openwrt

generic: track and understand wrong formatted patch in hack/pending #10272

THE LIST