What's the protocol for UAPI headers in staging drivers? Specifically, I'm wondering how people usually install those headers. For example, there are a bunch of Android UAPI headers under drivers/staging/android/uapi which currently end up at <INSTALL_HDR_PATH>/drivers with a call to `make headers_install'. All other UAPI headers end up at <INSTALL_HDR_PATH>/usr so getting at the staging ones requires some special handling in the "outer" build system... To get the staging UAPI headers installed to <INSTALL_HDR_PATH>/usr like the rest of the UAPI headers, we're currently running with something like this in our tree: --8<---------------cut here---------------start------------->8--- diff --git a/drivers/staging/android/uapi/Kbuild b/drivers/staging/android/uapi/Kbuild new file mode 100644 index 0000000000..c075619bf4 --- /dev/null +++ b/drivers/staging/android/uapi/Kbuild @@ -0,0 +1 @@ +header-y += ion.h diff --git a/include/uapi/linux/Kbuild b/include/uapi/linux/Kbuild index 6929571b79..8bba379192 100644 --- a/include/uapi/linux/Kbuild +++ b/include/uapi/linux/Kbuild @@ -20,6 +20,7 @@ header-y += netfilter_ipv4/ header-y += netfilter_ipv6/ header-y += usb/ header-y += wimax/ +header-y += ../../../drivers/staging/android/uapi/ genhdr-y += version.h diff --git a/scripts/Makefile.headersinst b/scripts/Makefile.headersinst index 8ccf83056a..79a64d7959 100644 --- a/scripts/Makefile.headersinst +++ b/scripts/Makefile.headersinst @@ -125,7 +125,7 @@ endif hdr-inst := -rR -f $(srctree)/scripts/Makefile.headersinst obj .PHONY: $(subdirs) $(subdirs): - $(Q)$(MAKE) $(hdr-inst)=$(obj)/$@ dst=$(_dst)/$@ + $(Q)$(MAKE) $(hdr-inst)=$(obj)/$@ dst=$(_dst)/$(subst drivers/staging,usr/include/linux/staging,$@) targets := $(wildcard $(sort $(targets))) cmd_files := $(wildcard \ --8<---------------cut here---------------end--------------->8--- With this, the headers end up in <INSTALL_HDR_PATH>/usr/include/linux/staging instead of <INSTALL_HDR_PATH>/drivers/staging. I know staging drivers are unstable but shouldn't they still be adhering to the UAPI split and shouldn't there be a way to install those split headers? How does everyone else deal with this issue? Thanks! Mitch -- The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, hosted by The Linux Foundation _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
