On 5/10/19 11:18 PM, Christian Lamparter wrote:

On Friday, May 10, 2019 3:56:37 PM CEST l...@allycomm.com wrote:
From: Jeff Kletsky <git-comm...@allycomm.com>

Incorporates multiple changes, including file-based input and output

 From upstream commit:

     commit 35386dfd13b7
     Author: Geert Uytterhoeven <geert+rene...@glider.be>
     Date:   Mon Sep 3 19:33:23 2018 +0200

Signed-off-by: Jeff Kletsky <git-comm...@allycomm.com>
---
Hm, this got me thinking. Because we could also just do what the perf
package does and built the spidev_test from the device's linux kernel
source and ship it. This would also entail that we don't need to ship
the spidev_test.c in package/utils/spidev_test/src anymore. What do you
think?

Cheers,
Christian

---
diff --git a/package/utils/spidev_test/Makefile 
b/package/utils/spidev_test/Makefile
index b7c5227f91..29c6a90dc8 100644
--- a/package/utils/spidev_test/Makefile
+++ b/package/utils/spidev_test/Makefile
@@ -10,13 +10,16 @@ include $(INCLUDE_DIR)/kernel.mk
PKG_NAME:=spidev-test
  PKG_RELEASE:=$(LINUX_VERSION)
+PKG_FLAGS:=nonshared
+PKG_BUILD_DIR:=$(LINUX_DIR)/tools/spi-$(TARGET_DIR_NAME)
+PKG_BUILD_PARALLEL:=1
include $(INCLUDE_DIR)/package.mk define Package/spidev-test
    SECTION:=utils
    CATEGORY:=Utilities
-  DEPENDS:=+kmod-spi-dev
+  DEPENDS:=+kmod-spi-dev @!IN_SDK
    TITLE:=SPI testing utility
    VERSION:=$(LINUX_VERSION)-$(PKG_RELEASE)
    URL:=http://www.kernel.org
@@ -27,9 +30,26 @@ define Package/spidev-test/description
    SPI testing utility.
  endef
+define Build/Prepare
+       $(CP) $(LINUX_DIR)/tools/spi/* $(PKG_BUILD_DIR)/
+endef

[...]

Sounds great to me!

It's one of those packages that probably only a couple people a year
need, but when they need it, they really need it. The way you proposed
means that it is as current as the Linux version in hand without further
attention.

I was going to say that it would be a couple days until I was back home
and able to look into it and test as I'm on travel, but it looks like
you've already got things well in order. If you want to go ahead with
your patch, I think it's a much better approach.

Jeff

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

Reply via email to