[OpenWrt-Devel] Build issue using ccache due to GCC patch / R_MIPS_26

2015-07-14 Thread Karl Vogel 
There seems to be an issue using ccache in combination with the openwrt 
patched gcc.


Found the issue while investigating the following compilation issue:

---
openwrt/staging_dir/toolchain-mips_gcc-4.6-linaro_uClibc-0.9.33.2/lib/gcc/mips-openwrt-linux-uclibc/4.6.3/../../../../mips-openwrt-linux-uclibc/bin/ld: 
.libs/libxt_ACCOUNT_cl.o: relocation R_MIPS_26 against `memset' can not 
be used when making a shared object; recompile with -fPIC

libs/libxt_ACCOUNT_cl.o: could not read symbols: Bad value
collect2: ld returned 1 exit status
make[8]: *** [libxt_ACCOUNT_cl.la] Error 1
---


This only happened sometimes and only on certain machines. Closer look 
showed that libtool wasn't actually adding -fPIC. This was caused by 
the autoconf check for the PIC flag sometimes failing on my machine:


  checking for ccache_cc option to produce PIC... -fPIC -DPIC
  checking if ccache_cc PIC flag -fPIC -DPIC works... no

Doing a 'clean prepare compile' of the package sometimes fixed it.

The problem is that the autoconf test is doing a compilation with -fPIC 
then compares the stdout of the compiler to a 'normal' compilation. The 
results have to be equal to pass the test, in my setup, this wasn't 
always the case.


The reason was that the compilation output for that step sometimes gave 
a warning:


  warning: someone does not honour COPTS correctly, passed 2 times

This warning comes from the 910-mbsd_multi.patch patch to gcc, which 
verifies that -fhonour-copts is on the compilation line and that it is 
only present once. The patch also uses an environment variable called 
GCC_HONOUR_COPTS which indicates what has to happen if the option isn't 
present or is specified more than once. (possible values 0, 1, 2 and 
's' for silent)



The problem that was occuring on my setup was that -fhonour-copts had 
been specified more than once for an autoconf check and that this 
result was cached by ccache.


All subsequent compilations for the same autoconf conftest.c file, 
returned the result from the compiler cache as for ccache the same 
inputs where used for the compilation. The problem here is that ccache 
has no knowledge that the environment variable GCC_HONOUR_COPTS 
influences the stdout from gcc and since the stdout is of importance to 
autoconf, it was causing the PIC config test to fail due to a prior 
cached compilation result.


To verify my finding, I tried compilation without ccache, which worked 
and then tried patching ccache to also track GCC_HONOUR_COPTS, which 
also produced the correct result:


--- ccache-3.1.7/ccache.c   2012-01-08 15:40:55.0 +0100
+++ ccache-3.1.7-patched/ccache.c   2015-07-14 09:56:02.037675777 
+0200

@@ -965,6 +965,7 @@ calculate_object_hash(struct args *args,
   "CPLUS_INCLUDE_PATH",
   "OBJC_INCLUDE_PATH",
   "OBJCPLUS_INCLUDE_PATH", /* clang */
+"GCC_HONOUR_COPTS",
   NULL
   };
   for (p = envvars; *p != NULL ; ++p) {

With this patch, ccache will also add the GCC_HONOUR_COPTS setting to 
the hash of the compilation command and will consider 2 compilations 
with a different GCC_HONOUR_COPTS to actually be 2 different 
compilations.



However, this doesn't completely fix the problem as it appears that the 
tools/ccache/Makefile will only be compiled if ccache isn't already 
present on the host system, so probably the proper way to fix this is 
to just unconditionally compile ccache too if it is selected by the 
build system and then add the above ccache patch to the package.


Regards,
Karl
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] [PATCH] Make ccache aware of OpenWRT provided GCC patch.

2015-07-14 Thread Karl Vogel 

The OpenWRT GCC patch, 910-mbsd_multi.patch adds an extra option
to gcc that depends on an environment variable. Standard ccache
is unaware of this option and therefor can produce stdout from gcc
that doesn't correspond to what you would get if you actually
do a direct compilation with gcc.

This commit adds a patch to ccache to make it aware of the new
option and removes the feature from the ccache package to use the
host system's installed ccache binary.
---
tools/ccache/Makefile   | 24 --
tools/ccache/patches/100-honour-copts.patch | 31 
+

2 files changed, 35 insertions(+), 20 deletions(-)
create mode 100644 tools/ccache/patches/100-honour-copts.patch

diff --git a/tools/ccache/Makefile b/tools/ccache/Makefile
index 9020c9c..75c5f8c 100644
--- a/tools/ccache/Makefile
+++ b/tools/ccache/Makefile
@@ -24,29 +24,13 @@ define Host/Install/ccache
$(CP) ./files/* $(STAGING_DIR_HOST)/bin/
endef

-ifneq ($(strip $(shell which ccache >/dev/null && echo found)),found)
-  define Host/Clean
+define Host/Clean
-$(MAKE) -C $(HOST_BUILD_DIR) uninstall
$(call Host/Clean/Default)
-  endef
-  define Host/Install
+endef
+define Host/Install
$(call Host/Install/Default)
$(call Host/Install/ccache)
-  endef
-else
-  define Host/Prepare
-  endef
-  define Host/Configure
-  endef
-  define Host/Compile
-  endef
-  define Host/Install
-   $(call Host/Install/ccache)
-  endef
-  define Host/Clean
-  endef
-  define Download
-  endef
-endif
+endef

$(eval $(call HostBuild))
diff --git a/tools/ccache/patches/100-honour-copts.patch 
b/tools/ccache/patches/100-honour-copts.patch

new file mode 100644
index 000..9182d33
--- /dev/null
+++ b/tools/ccache/patches/100-honour-copts.patch
@@ -0,0 +1,31 @@
+From 135d868c539f994afbcf80d313ba2e39f031d3b1 Mon Sep 17 00:00:00 2001
+From: Karl Vogel 
+Date: Tue, 14 Jul 2015 11:05:33 +0200
+Subject: [PATCH] Include environment variable GCC_HONOUR_COPTS in hash.
+
+The OpenWRT patch, 910-mbsd_multi.patch, to GCC adds an extra
+compilation flag, -fhonour-copts, which is influenced by an
+environment variable called GCC_HONOUR_COPTS.
+
+Include this environment var in the hash calculation as otherwise
+the gcc stdout warning from a previous compilation might be shown
+where, even when GCC_HONOUR_COPTS is in 's'ilent mode.
+---
+ ccache.c | 1 +
+ 1 file changed, 1 insertion(+)
+
+diff --git a/ccache.c b/ccache.c
+index e41af13..fa67ed2 100644
+--- a/ccache.c
 b/ccache.c
+@@ -965,6 +965,7 @@ calculate_object_hash(struct args *args, struct 
mdfour *hash, int direct_mode)

+   "CPLUS_INCLUDE_PATH",
+   "OBJC_INCLUDE_PATH",
+   "OBJCPLUS_INCLUDE_PATH", /* clang */
++"GCC_HONOUR_COPTS",
+   NULL
+   };
+   for (p = envvars; *p != NULL ; ++p) {
+--
+1.9.1
+
--
1.9.1
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] [PATCH] Make ccache aware of OpenWRT provided GCC patch.

2015-07-14 Thread Karl Vogel 
The OpenWRT GCC patch, 910-mbsd_multi.patch adds an extra option
to gcc that depends on an environment variable. Standard ccache
is unaware of this option and therefor can produce stdout from gcc
that doesn't correspond to what you would get if you actually
do a direct compilation with gcc.

This commit adds a patch to ccache to make it aware of the new
option and removes the feature from the ccache package to use the
host system's installed ccache binary.
---
 tools/ccache/Makefile   | 24 --
 tools/ccache/patches/100-honour-copts.patch | 31 +
 2 files changed, 35 insertions(+), 20 deletions(-)
 create mode 100644 tools/ccache/patches/100-honour-copts.patch

diff --git a/tools/ccache/Makefile b/tools/ccache/Makefile
index 9020c9c..75c5f8c 100644
--- a/tools/ccache/Makefile
+++ b/tools/ccache/Makefile
@@ -24,29 +24,13 @@ define Host/Install/ccache
$(CP) ./files/* $(STAGING_DIR_HOST)/bin/
 endef
 
-ifneq ($(strip $(shell which ccache >/dev/null && echo found)),found)
-  define Host/Clean
+define Host/Clean
-$(MAKE) -C $(HOST_BUILD_DIR) uninstall
$(call Host/Clean/Default)
-  endef
-  define Host/Install
+endef
+define Host/Install
$(call Host/Install/Default)
$(call Host/Install/ccache)
-  endef
-else
-  define Host/Prepare
-  endef
-  define Host/Configure
-  endef
-  define Host/Compile
-  endef
-  define Host/Install
-   $(call Host/Install/ccache)
-  endef
-  define Host/Clean
-  endef
-  define Download
-  endef
-endif
+endef
 
 $(eval $(call HostBuild))
diff --git a/tools/ccache/patches/100-honour-copts.patch 
b/tools/ccache/patches/100-honour-copts.patch
new file mode 100644
index 000..9182d33
--- /dev/null
+++ b/tools/ccache/patches/100-honour-copts.patch
@@ -0,0 +1,31 @@
+From 135d868c539f994afbcf80d313ba2e39f031d3b1 Mon Sep 17 00:00:00 2001
+From: Karl Vogel 
+Date: Tue, 14 Jul 2015 11:05:33 +0200
+Subject: [PATCH] Include environment variable GCC_HONOUR_COPTS in hash.
+
+The OpenWRT patch, 910-mbsd_multi.patch, to GCC adds an extra
+compilation flag, -fhonour-copts, which is influenced by an
+environment variable called GCC_HONOUR_COPTS.
+
+Include this environment var in the hash calculation as otherwise
+the gcc stdout warning from a previous compilation might be shown
+where, even when GCC_HONOUR_COPTS is in 's'ilent mode.
+---
+ ccache.c | 1 +
+ 1 file changed, 1 insertion(+)
+
+diff --git a/ccache.c b/ccache.c
+index e41af13..fa67ed2 100644
+--- a/ccache.c
 b/ccache.c
+@@ -965,6 +965,7 @@ calculate_object_hash(struct args *args, struct mdfour 
*hash, int direct_mode)
+   "CPLUS_INCLUDE_PATH",
+   "OBJC_INCLUDE_PATH",
+   "OBJCPLUS_INCLUDE_PATH", /* clang */
++"GCC_HONOUR_COPTS",
+   NULL
+   };
+   for (p = envvars; *p != NULL ; ++p) {
+-- 
+1.9.1
+
-- 
1.9.1
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] [PATCH] ccache: Make ccache aware of OpenWRT provided GCC patch.

2015-07-14 Thread Karl Vogel 
The OpenWRT GCC patch, 910-mbsd_multi.patch adds an extra option
to gcc that depends on an environment variable. Standard ccache
is unaware of this option and therefor can produce stdout from gcc
that doesn't correspond to what you would get if you actually
do a direct compilation with gcc.

This commit adds a patch to ccache to make it aware of the new
option and removes the feature from the ccache package to use the
host system's installed ccache binary.

Signed-off-by: Karl Vogel 
---
 tools/ccache/Makefile   | 24 -
 tools/ccache/patches/100-honour-copts.patch | 33 +
 2 files changed, 37 insertions(+), 20 deletions(-)
 create mode 100644 tools/ccache/patches/100-honour-copts.patch

diff --git a/tools/ccache/Makefile b/tools/ccache/Makefile
index 9020c9c..75c5f8c 100644
--- a/tools/ccache/Makefile
+++ b/tools/ccache/Makefile
@@ -24,29 +24,13 @@ define Host/Install/ccache
$(CP) ./files/* $(STAGING_DIR_HOST)/bin/
 endef
 
-ifneq ($(strip $(shell which ccache >/dev/null && echo found)),found)
-  define Host/Clean
+define Host/Clean
-$(MAKE) -C $(HOST_BUILD_DIR) uninstall
$(call Host/Clean/Default)
-  endef
-  define Host/Install
+endef
+define Host/Install
$(call Host/Install/Default)
$(call Host/Install/ccache)
-  endef
-else
-  define Host/Prepare
-  endef
-  define Host/Configure
-  endef
-  define Host/Compile
-  endef
-  define Host/Install
-   $(call Host/Install/ccache)
-  endef
-  define Host/Clean
-  endef
-  define Download
-  endef
-endif
+endef
 
 $(eval $(call HostBuild))
diff --git a/tools/ccache/patches/100-honour-copts.patch 
b/tools/ccache/patches/100-honour-copts.patch
new file mode 100644
index 000..5496dc1
--- /dev/null
+++ b/tools/ccache/patches/100-honour-copts.patch
@@ -0,0 +1,33 @@
+From e9a75b66ebb0b5b289b69a188b8b16306151bd11 Mon Sep 17 00:00:00 2001
+From: Karl Vogel 
+Date: Tue, 14 Jul 2015 11:05:33 +0200
+Subject: [PATCH] Include environment variable GCC_HONOUR_COPTS in hash.
+
+The OpenWRT patch, 910-mbsd_multi.patch, to GCC adds an extra
+compilation flag, -fhonour-copts, which is influenced by an
+environment variable called GCC_HONOUR_COPTS.
+
+Include this environment var in the hash calculation as otherwise
+the gcc stdout warning from a previous compilation might be shown
+where, even when GCC_HONOUR_COPTS is in 's'ilent mode.
+
+Signed-off-by: Karl Vogel 
+---
+ ccache.c | 1 +
+ 1 file changed, 1 insertion(+)
+
+diff --git a/ccache.c b/ccache.c
+index e41af13..fa67ed2 100644
+--- a/ccache.c
 b/ccache.c
+@@ -965,6 +965,7 @@ calculate_object_hash(struct args *args, struct mdfour 
*hash, int direct_mode)
+   "CPLUS_INCLUDE_PATH",
+   "OBJC_INCLUDE_PATH",
+   "OBJCPLUS_INCLUDE_PATH", /* clang */
++"GCC_HONOUR_COPTS",
+   NULL
+   };
+   for (p = envvars; *p != NULL ; ++p) {
+-- 
+1.9.1
+
-- 
1.9.1
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] [PATCH] ccache: Make ccache aware of OpenWRT provided GCC patch.

2015-07-14 Thread Karl Vogel 
The OpenWRT GCC patch, 910-mbsd_multi.patch adds an extra option
to gcc that depends on an environment variable. Standard ccache
is unaware of this option and therefor can produce stdout from gcc
that doesn't correspond to what you would get if you actually
do a direct compilation with gcc.

This commit adds a patch to ccache to make it aware of the new
option and removes the feature from the ccache package to use the
host system's installed ccache binary.

Signed-off-by: Karl Vogel 
---
 tools/ccache/Makefile   | 24 -
 tools/ccache/patches/100-honour-copts.patch | 33 +
 2 files changed, 37 insertions(+), 20 deletions(-)
 create mode 100644 tools/ccache/patches/100-honour-copts.patch

diff --git a/tools/ccache/Makefile b/tools/ccache/Makefile
index 9020c9c..75c5f8c 100644
--- a/tools/ccache/Makefile
+++ b/tools/ccache/Makefile
@@ -24,29 +24,13 @@ define Host/Install/ccache
$(CP) ./files/* $(STAGING_DIR_HOST)/bin/
 endef
 
-ifneq ($(strip $(shell which ccache >/dev/null && echo found)),found)
-  define Host/Clean
+define Host/Clean
-$(MAKE) -C $(HOST_BUILD_DIR) uninstall
$(call Host/Clean/Default)
-  endef
-  define Host/Install
+endef
+define Host/Install
$(call Host/Install/Default)
$(call Host/Install/ccache)
-  endef
-else
-  define Host/Prepare
-  endef
-  define Host/Configure
-  endef
-  define Host/Compile
-  endef
-  define Host/Install
-   $(call Host/Install/ccache)
-  endef
-  define Host/Clean
-  endef
-  define Download
-  endef
-endif
+endef
 
 $(eval $(call HostBuild))
diff --git a/tools/ccache/patches/100-honour-copts.patch 
b/tools/ccache/patches/100-honour-copts.patch
new file mode 100644
index 000..dc32885
--- /dev/null
+++ b/tools/ccache/patches/100-honour-copts.patch
@@ -0,0 +1,33 @@
+From 90762a9b8d9a50b6176f10bd6c2e2b9501117561 Mon Sep 17 00:00:00 2001
+From: Karl Vogel 
+Date: Tue, 14 Jul 2015 11:05:33 +0200
+Subject: [PATCH] Include environment variable GCC_HONOUR_COPTS in hash.
+
+The OpenWRT patch, 910-mbsd_multi.patch, to GCC adds an extra
+compilation flag, -fhonour-copts, which is influenced by an
+environment variable called GCC_HONOUR_COPTS.
+
+Include this environment var in the hash calculation as otherwise
+the gcc stdout warning from a previous compilation might be shown
+where, even when GCC_HONOUR_COPTS is in 's'ilent mode.
+
+Signed-off-by: Karl Vogel 
+---
+ ccache.c | 1 +
+ 1 file changed, 1 insertion(+)
+
+diff --git a/ccache.c b/ccache.c
+index e41af13..b736a9c 100644
+--- a/ccache.c
 b/ccache.c
+@@ -965,6 +965,7 @@ calculate_object_hash(struct args *args, struct mdfour 
*hash, int direct_mode)
+   "CPLUS_INCLUDE_PATH",
+   "OBJC_INCLUDE_PATH",
+   "OBJCPLUS_INCLUDE_PATH", /* clang */
++  "GCC_HONOUR_COPTS",
+   NULL
+   };
+   for (p = envvars; *p != NULL ; ++p) {
+-- 
+1.9.1
+
-- 
1.9.1
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] fstools: MTD driver - deadc0de marker

2016-03-02 Thread Karl Vogel 
I ran into an issue on a platform using a JFFS2 partition where the
overlayfs mount was done with tmpfs. After some investigation the issue
seemed to happen in the mount_root binary, more specifically during the
mtd_volume_identify() function.

The code does the check:

  if (v->type == UBIVOLUME && deadc0de == 0x) {
  return FS_JFFS2;
  }

see
http://git.openwrt.org/project/fstools.git?p=project/fstools.git;a=blob;f=libfstools/mtd.c;h=156d166b5a2cc92f124852180782e5c2826be881;hb=HEAD#l215


However, in my case the volume type is NANDFLASH. Adding a v->type ==
NANDFLASH in this place, makes my system work, but I'm not entirely sure
why this deadc0de check was restricted to UBIVOLUME only?! Would a check
for NANDFLASH be acceptable here?

Regards,
Karl
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] Read-only mirror https://github.com/openwrt/openwrt is stuck since 12 days ago

2016-03-16 Thread Karl Vogel 
On Tue, Mar 15, 2016 at 6:22 PM, Bruno Randolf  wrote:

> On 03/07/2016 10:51 AM, John Crispin wrote:
> >
> >
> > On 07/03/2016 11:47, Álvaro Fernández Rojas wrote:
> >> I just want to add that openwrt-commits list is also affected, since
> last commit published there matches the last one in the read-only mirror.
> >>
> >> Regards,
> >> Álvaro.
> >
> > i noticed that one aswell, will investigate ... sorry for the
> > delay/inconvenience
>
> It is still down:
>
> https://github.com/openwrt/openwrt?
> "Updated 21 days ago"...
>
> Would be great to have the git repo synced to svn again.
>
> Thanks,
> bruno
>

Doesn't look like the SVN server is back yet either.

Are there any plans to resurrect the anon SVN server? Or should everybody
move to git.openwrt.org instead?



> >>> El 6 mar 2016, a las 19:33, Hannu Nyman  escribió:
> >>>
> >>> The read-only mirror for the main Openwrt sources at Github has been
> stuck since 12 days ago.
> >>>
> >>> https://github.com/openwrt/openwrt
> >>>
> >>> Both trunk and Chaos Calmer have got new commits since then...
> >>>
> >>> Apparently the mirror has not returned to automatic updates after the
> network/hardware troubles.
> ___
> openwrt-devel mailing list
> openwrt-devel@lists.openwrt.org
> https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
>
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] [PATCH] toolchain/kernel-headers: use target config when generating headers

2016-03-24 Thread Karl Vogel 
Without this change, the oldconfig could actually use the host
system's kernel defaults from /boot, ie on an ubuntu build node
the oldconfig would give:

 # using defaults found in /boot/config-3.13.0-77-generic

This patch uses the Kernel/Prepare and Kernel/Configure for the
actual target to generate the .config. While the kernel to userspace
ABI is fixed, there might be additions to the kernel which would
not be known when using the host configuration.

ie. in the past the iptables Makefile was altered because the
toolchain kernel headers were actually missing the layer7 headers
(layer7 patches has since been dropped), but it indicates the
issue.

With this patch, the toolchain kernel headers will be those
of the actual target kernel.

Signed-off-by: Karl Vogel 
---
 include/kernel-defaults.mk| 2 ++
 toolchain/kernel-headers/Makefile | 7 +++
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/include/kernel-defaults.mk b/include/kernel-defaults.mk
index 406fd46..19ff994 100644
--- a/include/kernel-defaults.mk
+++ b/include/kernel-defaults.mk
@@ -66,6 +66,7 @@ else
ln -s $(CONFIG_EXTERNAL_KERNEL_TREE) $(LINUX_DIR)
   endef
 endif
+Kernel/Prepare?=$(Kernel/Prepare/Default)
 
 ifeq ($(CONFIG_TARGET_ROOTFS_INITRAMFS),y)
   ifeq ($(strip $(CONFIG_EXTERNAL_CPIO)),"")
@@ -122,6 +123,7 @@ define Kernel/Configure/Default
$(_SINGLE) [ -d $(LINUX_DIR)/user_headers ] || $(MAKE) 
$(KERNEL_MAKEOPTS) INSTALL_HDR_PATH=$(LINUX_DIR)/user_headers headers_install
$(SH_FUNC) grep '=[ym]' $(LINUX_DIR)/.config.set | LC_ALL=C sort | md5s 
> $(LINUX_DIR)/.vermagic
 endef
+Kernel/Configure?=$(Kernel/Configure/Default)
 
 define Kernel/Configure/Initramfs
$(call Kernel/SetInitramfs)
diff --git a/toolchain/kernel-headers/Makefile 
b/toolchain/kernel-headers/Makefile
index 68c83a0..fae20d4 100644
--- a/toolchain/kernel-headers/Makefile
+++ b/toolchain/kernel-headers/Makefile
@@ -46,9 +46,7 @@ KMAKE := $(MAKE) -C $(HOST_BUILD_DIR) \
 
 define Host/Configure/all
mkdir -p $(BUILD_DIR_TOOLCHAIN)/linux-dev
-   $(KMAKE) \
-   INSTALL_HDR_PATH="$(BUILD_DIR_TOOLCHAIN)/linux-dev/" \
-   headers_install
+   $(CP) $(LINUX_DIR)/user_headers/* $(BUILD_DIR_TOOLCHAIN)/linux-dev/
 endef
 
 # XXX: the following is needed to build lzma-loader
@@ -71,12 +69,13 @@ define Host/Configure/post/mipsel
 endef
 
 define Host/Prepare
-   $(call Kernel/Prepare/Default)
+   $(call Kernel/Prepare)
ln -sf linux-$(LINUX_VERSION) $(BUILD_DIR_TOOLCHAIN)/linux
$(SED) 's/@expr length/@-expr length/' $(HOST_BUILD_DIR)/Makefile
 endef
 
 define Host/Configure
+   $(call Kernel/Configure)
env
yes '' | $(KMAKE) oldconfig
$(call Host/Configure/all)
-- 
2.7.4
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] [PATCH] ubus: Correct usage of timeout on poll function

2015-12-06 Thread Karl Vogel 
As the man page explains:

"Specifying a timeout of zero causes poll() to return immediately,
even if no file descriptors are ready."

The use of 0 as timeout could cause libubus to busy loop if the
socket was non-blocking. For blocking sockets, this was less
apparent as the subsequent read() would then block.

Signed-off-by: Karl Vogel 
---
 libubus-io.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/libubus-io.c b/libubus-io.c
index 9d3ac6c..e6d4236 100644
--- a/libubus-io.c
+++ b/libubus-io.c
@@ -54,7 +54,7 @@ static void wait_data(int fd, bool write)
struct pollfd pfd = { .fd = fd };
 
pfd.events = write ? POLLOUT : POLLIN;
-   poll(&pfd, 1, 0);
+   poll(&pfd, 1, -1);
 }
 
 static int writev_retry(int fd, struct iovec *iov, int iov_len, int sock_fd)
@@ -321,7 +321,7 @@ void __hidden ubus_poll_data(struct ubus_context *ctx, int 
timeout)
.events = POLLIN | POLLERR,
};
 
-   poll(&pfd, 1, timeout);
+   poll(&pfd, 1, timeout ? timeout : -1);
ubus_handle_data(&ctx->sock, ULOOP_READ);
 }
 
-- 
2.6.3
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] Mismatch in ubus / uloop options

2015-12-07 Thread Karl Vogel 
In ubus_reconnect, the ctx->sock.fd is created and set to non-blocking mode:

fcntl(ctx->sock.fd, F_SETFL, fcntl(ctx->sock.fd, F_GETFL) | O_NONBLOCK |
> O_CLOEXEC);
>

http://git.openwrt.org/?p=project/ubus.git;a=blob;f=libubus-io.c;h=9d3ac6c33f4d926c7f5c5d2033d531fc5fc6f834;hb=259450f414d8c9ee41896e8e6d6bc57ec00e2b63#l396


However, further in the call stack, the socket will get added to uloop via
ubus_add_uloop with the flag ULOOP_BLOCKING set.

http://git.openwrt.org/?p=project/ubus.git;a=blob;f=libubus.h;h=54e7d481b0496e719f2f567b664a2619c82a3b16;hb=259450f414d8c9ee41896e8e6d6bc57ec00e2b63#l245

ie when called from ubus_complete_request on an !registered ctx->sock, the
ubus_add_uloop is done:

http://git.openwrt.org/?p=project/ubus.git;a=blob;f=libubus-req.c;h=bf19f3670ea5b215190dfcce125c358f616c72a4;hb=259450f414d8c9ee41896e8e6d6bc57ec00e2b63#l136

The uloop function assumes a socket is blocking by default and will never
make a socket blocking again, it will only turn it to non-blocking if
ULOOP_BLOCKING isn't set. This caused a busy loop in some daemons that had
their socket set to non-blocking due to another issue, ie. poll using 0 as
infinite instead of -1, for which I submitted a fix earlier.

I wonder if it wouldn't be better to have uloop set the socket flags in all
cases? Currently it "assumes" the socket is blocking and if ULOOP_BLOCKING
is passed, it will set it to non-blocking, but it won't do the other way
round. Though that might have some impact on daemons that think the socket
is non-blocking while it will be turned into blocking if this is changed.
So not sure what the best way to go about this one is?!


Regards,
Karl
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] network/znc: Package binaries appear to be broken by MIPS16

2015-01-21 Thread Karl Vogel 
On Tue, Jan 20, 2015 at 11:07 AM, Sławomir Demeszko
 wrote:
> 19.01.2015 19:50, Oliver wrote:
>>
>> I'll try 4.9 - out of interest, do you have any references to the bug in
>> 4.8 (C++ specific?)
>
>
> Sorry, I do not remember where I read about that.

I think you're referring to the MIPS16 bug in qemu:

http://wiki.openwrt.org/doc/howto/qemu#openwrt_in_qemu_mips
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] [PATCH][packages] Enable better customization of nginx features.

2012-11-16 Thread Karl Vogel 
Add configuration settings to disable certain features in nginx to
be able to reduce it's footprint and allow nginx to be built with
support for lua.

Signed-off-by: Karl Vogel 
---
 net/nginx/Config.in   |  131 +
 net/nginx/Makefile|  125 +++-
 net/nginx/patches-lua-nginx/300-ldl.patch |   23 +
 3 files changed, 277 insertions(+), 2 deletions(-)
 create mode 100644 net/nginx/patches-lua-nginx/300-ldl.patch

diff --git a/net/nginx/Config.in b/net/nginx/Config.in
index e451037..1396d10 100644
--- a/net/nginx/Config.in
+++ b/net/nginx/Config.in
@@ -36,4 +36,135 @@ config NGINX_DAV
help
Enable the HTTP and WebDAV methods PUT, DELETE, MKCOL, COPY and 
MOVE.
 
+config NGINX_MODULE_LUA
+   bool
+   prompt "Enable LUA module"
+   default n
+   help
+   Enable support for LUA scripts.
+
+   menu "Disable features"
+
+   config NGINX_DISABLE_PCRE
+ bool
+ prompt "Disable PCRE library usage"
+ default n
+
+   config NGINX_DISABLE_HTTP_CACHE
+ depends on NGINX_SSL=y
+ bool
+ prompt "Disable HTTP cache"
+ default y
+
+   config NGINX_DISABLE_HTTP_CHARSET_MODULE
+ bool
+ prompt "Disable HTTP charset module"
+ default n
+
+   config NGINX_DISABLE_HTTP_GZIP_MODULE
+ bool
+ prompt "Disable HTTP gzip module"
+ default n
+
+   config NGINX_DISABLE_HTTP_SSI_MODULE
+ bool
+ prompt "Disable HTTP ssi module"
+ default n
+
+   config NGINX_DISABLE_HTTP_USERID_MODULE
+ bool
+ prompt "Disable HTTP userid module"
+ default n
+
+   config NGINX_DISABLE_HTTP_ACCESS_MODULE
+ bool
+ prompt "Disable HTTP access module"
+ default n
+
+   config NGINX_DISABLE_HTTP_AUTH_BASIC_MODULE
+ bool
+ prompt "Disable HTTP auth basic"
+ default n
+
+   config NGINX_DISABLE_HTTP_AUTOINDEX_MODULE
+ bool
+ prompt "Disable HTTP autoindex module"
+ default n
+
+   config NGINX_DISABLE_HTTP_GEO_MODULE
+ bool
+ prompt "Disable HTTP geo module"
+ default n
+
+   config NGINX_DISABLE_HTTP_MAP_MODULE
+ bool
+ prompt "Disable HTTP map module"
+ default n
+
+   config NGINX_DISABLE_HTTP_SPLIT_CLIENTS_MODULE
+ bool
+ prompt "Disable HTTP split clients"
+ default n
+
+   config NGINX_DISABLE_HTTP_REFERER_MODULE
+ bool
+ prompt "Disable HTTP referer module"
+ default n
+
+   config NGINX_DISABLE_HTTP_REWRITE_MODULE
+ bool
+ prompt "Disable HTTP rewrite module"
+ default n
+
+   config NGINX_DISABLE_HTTP_PROXY_MODULE
+ bool
+ prompt "Disable HTTP proxy module"
+ default n
+
+   config NGINX_DISABLE_HTTP_FASTCGI_MODULE
+ bool
+ prompt "Disable HTTP fastcgi module"
+ default n
+
+   config NGINX_DISABLE_HTTP_UWSGI_MODULE
+ bool
+ prompt "Disable HTTP uwsgi module"
+ default n
+
+   config NGINX_DISABLE_HTTP_SCGI_MODULE
+ bool
+ prompt "Disable HTTP scgi module"
+ default n
+
+   config NGINX_DISABLE_HTTP_MEMCACHED_MODULE
+ bool
+ prompt "Disable HTTP memcached module"
+ default n
+
+   config NGINX_DISABLE_HTTP_LIMIT_CONN_MODULE
+ bool
+ prompt "Disable HTTP limit conn"
+ default n
+
+   config NGINX_DISABLE_HTTP_LIMIT_REQ_MODULE
+ bool
+ prompt "Disable HTTP limit req"
+ default n
+
+   config NGINX_DISABLE_HTTP_EMPTY_GIF_MODULE
+ bool
+ prompt "Disable HTTP empty gif"
+ default n
+
+   config NGINX_DISABLE_HTTP_BROWSER_MODULE
+ bool
+ prompt "Disable HTTP browser module"
+ default n
+
+   config NGINX_DISABLE_HTTP_UPSTREAM_IP_HASH_MODULE
+ bool
+ prompt "Disable HTTP IP hash module"
+ default n
+
+   endmenu
 endmenu
diff --git a/net/nginx/Makefile b/net/nginx/Makefile
index d56a729..e6fcbc8 100644
--- a/net/nginx/Makefile
+++ b/net/nginx/Makefile
@@ -22,7 +22,31 @@ PKG_CONFIG_DEPENDS := \
CONFIG_NGINX_STUB_STATUS \
CONFIG_NGINX_FLV \
CONFIG_NGINX_SSL \

[OpenWrt-Devel] [PATCH][packages] Enable better customization of nginx features.

2012-11-16 Thread Karl Vogel 
Add configuration settings to disable certain features in nginx to
be able to reduce it's footprint and allow nginx to be built with
support for lua.

Signed-off-by: Karl Vogel 
---
 net/nginx/Config.in   |  131 +
 net/nginx/Makefile|  125 +++-
 net/nginx/patches-lua-nginx/300-ldl.patch |   23 +
 3 files changed, 277 insertions(+), 2 deletions(-)
 create mode 100644 net/nginx/patches-lua-nginx/300-ldl.patch

diff --git a/net/nginx/Config.in b/net/nginx/Config.in
index e451037..1396d10 100644
--- a/net/nginx/Config.in
+++ b/net/nginx/Config.in
@@ -36,4 +36,135 @@ config NGINX_DAV
help
Enable the HTTP and WebDAV methods PUT, DELETE, MKCOL, COPY and 
MOVE.
 
+config NGINX_MODULE_LUA
+   bool
+   prompt "Enable LUA module"
+   default n
+   help
+   Enable support for LUA scripts.
+
+   menu "Disable features"
+
+   config NGINX_DISABLE_PCRE
+ bool
+ prompt "Disable PCRE library usage"
+ default n
+
+   config NGINX_DISABLE_HTTP_CACHE
+ depends on NGINX_SSL=y
+ bool
+ prompt "Disable HTTP cache"
+ default y
+
+   config NGINX_DISABLE_HTTP_CHARSET_MODULE
+ bool
+ prompt "Disable HTTP charset module"
+ default n
+
+   config NGINX_DISABLE_HTTP_GZIP_MODULE
+ bool
+ prompt "Disable HTTP gzip module"
+ default n
+
+   config NGINX_DISABLE_HTTP_SSI_MODULE
+ bool
+ prompt "Disable HTTP ssi module"
+ default n
+
+   config NGINX_DISABLE_HTTP_USERID_MODULE
+ bool
+ prompt "Disable HTTP userid module"
+ default n
+
+   config NGINX_DISABLE_HTTP_ACCESS_MODULE
+ bool
+ prompt "Disable HTTP access module"
+ default n
+
+   config NGINX_DISABLE_HTTP_AUTH_BASIC_MODULE
+ bool
+ prompt "Disable HTTP auth basic"
+ default n
+
+   config NGINX_DISABLE_HTTP_AUTOINDEX_MODULE
+ bool
+ prompt "Disable HTTP autoindex module"
+ default n
+
+   config NGINX_DISABLE_HTTP_GEO_MODULE
+ bool
+ prompt "Disable HTTP geo module"
+ default n
+
+   config NGINX_DISABLE_HTTP_MAP_MODULE
+ bool
+ prompt "Disable HTTP map module"
+ default n
+
+   config NGINX_DISABLE_HTTP_SPLIT_CLIENTS_MODULE
+ bool
+ prompt "Disable HTTP split clients"
+ default n
+
+   config NGINX_DISABLE_HTTP_REFERER_MODULE
+ bool
+ prompt "Disable HTTP referer module"
+ default n
+
+   config NGINX_DISABLE_HTTP_REWRITE_MODULE
+ bool
+ prompt "Disable HTTP rewrite module"
+ default n
+
+   config NGINX_DISABLE_HTTP_PROXY_MODULE
+ bool
+ prompt "Disable HTTP proxy module"
+ default n
+
+   config NGINX_DISABLE_HTTP_FASTCGI_MODULE
+ bool
+ prompt "Disable HTTP fastcgi module"
+ default n
+
+   config NGINX_DISABLE_HTTP_UWSGI_MODULE
+ bool
+ prompt "Disable HTTP uwsgi module"
+ default n
+
+   config NGINX_DISABLE_HTTP_SCGI_MODULE
+ bool
+ prompt "Disable HTTP scgi module"
+ default n
+
+   config NGINX_DISABLE_HTTP_MEMCACHED_MODULE
+ bool
+ prompt "Disable HTTP memcached module"
+ default n
+
+   config NGINX_DISABLE_HTTP_LIMIT_CONN_MODULE
+ bool
+ prompt "Disable HTTP limit conn"
+ default n
+
+   config NGINX_DISABLE_HTTP_LIMIT_REQ_MODULE
+ bool
+ prompt "Disable HTTP limit req"
+ default n
+
+   config NGINX_DISABLE_HTTP_EMPTY_GIF_MODULE
+ bool
+ prompt "Disable HTTP empty gif"
+ default n
+
+   config NGINX_DISABLE_HTTP_BROWSER_MODULE
+ bool
+ prompt "Disable HTTP browser module"
+ default n
+
+   config NGINX_DISABLE_HTTP_UPSTREAM_IP_HASH_MODULE
+ bool
+ prompt "Disable HTTP IP hash module"
+ default n
+
+   endmenu
 endmenu
diff --git a/net/nginx/Makefile b/net/nginx/Makefile
index d56a729..e6fcbc8 100644
--- a/net/nginx/Makefile
+++ b/net/nginx/Makefile
@@ -22,7 +22,31 @@ PKG_CONFIG_DEPENDS := \
CONFIG_NGINX_STUB_STATUS \
CONFIG_NGINX_FLV \
CONFIG_NGINX_SSL \

Re: [OpenWrt-Devel] [PATCH][packages] Enable better customization of nginx features.

2012-11-16 Thread Karl Vogel 
On Nov 16, 2012 11:16 AM, "Karl Vogel"  wrote:
>
> Add configuration settings to disable certain features in nginx to
> be able to reduce it's footprint and allow nginx to be built with
> support for lua.
>
> Signed-off-by: Karl Vogel 

Sorry for the duplicate spam.
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] [PATCH] ubus: Fix imbalance in lua stack push/pop of values.

2014-02-11 Thread Karl Vogel 
The lua getglobal and rawgeti both push a value onto the lua stack,
but they weren't being removed by the ubus_method_handler function,
thus corrupting the lua stack.

In case the specified method wasn't a function, the stack was also
corrupted as the method name remained on the stack.

Signed-off-by: Karl Vogel 
Cc: Felix Fietkau 
---
 lua/ubus.c |6 +-
 1 files changed, 5 insertions(+), 1 deletions(-)

diff --git a/lua/ubus.c b/lua/ubus.c
index 6da935d..0f2338c 100644
--- a/lua/ubus.c
+++ b/lua/ubus.c
@@ -289,6 +289,8 @@ ubus_method_handler(struct ubus_context *ctx, struct 
ubus_object *obj,
lua_getglobal(state, "__ubus_cb");
lua_rawgeti(state, -1, o->r);
lua_getfield(state, -1, method);
+   lua_remove(state, -2);
+   lua_remove(state, -2);
 
if (lua_isfunction(state, -1)) {
lua_pushlightuserdata(state, req);
@@ -297,7 +299,9 @@ ubus_method_handler(struct ubus_context *ctx, struct 
ubus_object *obj,
else
ubus_lua_parse_blob_array(state, blob_data(msg), 
blob_len(msg), true);
lua_call(state, 2, 0);
-   }
+   } else
+   lua_pop(state, 1);
+
return 0;
 }
 
-- 
1.7.1
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] [PATCH] uloop: Add flag to allow callback to be called on error conditions.

2014-02-11 Thread Karl Vogel 
In some conditions, an application is interested in errors happening
on a file descriptor and might be able to resolve the issue in the
callback function.

This patch adds a flag to notify the uloop framework that errors
should be passed to the callback function, instead of silently 
removing the fd from the polling set.

Signed-off-by: Karl Vogel 
Cc: Felix Fietkau 
---
 uloop.c |6 --
 uloop.h |2 ++
 2 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/uloop.c b/uloop.c
index 0566d80..d293edb 100644
--- a/uloop.c
+++ b/uloop.c
@@ -174,7 +174,8 @@ static int uloop_fetch_events(int timeout)
 
if (events[n].flags & EV_ERROR) {
u->error = true;
-   uloop_fd_delete(u);
+   if (!(u->flags & ULOOP_ERROR_CB))
+   uloop_fd_delete(u);
}
 
if(events[n].filter == EVFILT_READ)
@@ -268,7 +269,8 @@ static int uloop_fetch_events(int timeout)
 
if (events[n].events & (EPOLLERR|EPOLLHUP)) {
u->error = true;
-   uloop_fd_delete(u);
+   if (!(u->flags & ULOOP_ERROR_CB))
+   uloop_fd_delete(u);
}
 
if(!(events[n].events & 
(EPOLLRDHUP|EPOLLIN|EPOLLOUT|EPOLLERR|EPOLLHUP))) {
diff --git a/uloop.h b/uloop.h
index 98dd818..7564514 100644
--- a/uloop.h
+++ b/uloop.h
@@ -53,6 +53,8 @@ typedef void (*uloop_process_handler)(struct uloop_process 
*c, int ret);
 #define ULOOP_EDGE_DEFER   (1 << 5)
 #endif
 
+#define ULOOP_ERROR_CB (1 << 6)
+
 struct uloop_fd
 {
uloop_fd_handler cb;
-- 
1.7.1
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] [PATCH 1/5] netifd : Disable netlink auto ack

2014-02-11 Thread Karl Vogel 
On Tue, Feb 11, 2014 at 10:29:36AM +0100, Hans Dedecker wrote:
> Disable netlink auto ack when doing a delete in the get callback
> handler to avoid race conditions resulting into stalled message
> on the netlink socket.
> 
> Solves issue reported in https://dev.openwrt.org/ticket/14590
> 
> Signed-off-by: Karl Vogel 
> Acked-by: Hans Dedecker 
> ---
>  system-linux.c |5 +++--
>  1 files changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/system-linux.c b/system-linux.c
> index db78240..e1b9924 100644
> --- a/system-linux.c
> +++ b/system-linux.c
> @@ -456,8 +456,9 @@ static int cb_clear_event(struct nl_msg *msg, void *arg)
>   hdr->nlmsg_type = type;
>   hdr->nlmsg_flags = NLM_F_REQUEST;
>  
> - if (!nl_send_auto_complete(sock_rtnl, clr->msg))
> - nl_wait_for_ack(sock_rtnl);
> + nl_socket_disable_auto_ack(sock_rtnl);
> + nl_send_auto_complete(sock_rtnl, clr->msg);
> + nl_socket_enable_auto_ack(sock_rtnl);
>  
>   return NL_SKIP;
>  }
> -- 
> 1.7.1
> 

Just some more background information regarding this..

The current code had 2 issues. First issue was an incorrect return
code check, nl_send_auto_complete() returns the number of bytes
send or a negative value for an error. Therefor the wait for
ack nl_wait_for_ack() was never called in the original code.

The second issue is that the incoming netlink message is a
multipart message, so if the nl_wait_for_ack() was called, it
would get the next part of the multipart message, instead of
the actual ACK.

Since the original code never called the wait for ack, it didn't
run into this issue, but it did receive a stray ACK further down
the road, which caused other issues (like the netlink socket
not being totally read out, hence why the /proc/net/netlink Rmem
column showed outstanding buffered message on the socket).

This patch disables the auto ack generation for the delete, since
the ack isn't really useful (what would we do if the delete fails
anyway?). The only other way to work around this, would be to use
another netlink socket to use in the callback, but that requires
alot more resources.
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

Re: [OpenWrt-Devel] ubus_invoke & interruptions

2014-03-18 Thread Karl Vogel 
On Tue, Mar 18, 2014 at 4:06 PM, Felix Fietkau  wrote:
>
> On 2014-03-17 16:49, Helmut Schaa wrote:
> > On Mon, Mar 17, 2014 at 4:26 PM, Felix Fietkau  wrote:
> >> On 2014-03-17 09:18, Helmut Schaa wrote:
> >>> just realized that using ubus_invoke can get interrupted by other 
> >>> uloop-events
> >>> happening while executing a ubus method.
> >>>
> >>> Is this intended? Would it make sense to at least postpone all piled up 
> >>> events
> >>> to be executed "later"?
> >> Yes, this is intended. Postponing events in uloop would be a lot more
> >> complex than doing the same in the application using it.
> >
> > Fair enough. Being unaware of this can cause funny effects :D
> >
> >> uloop would have to handle blocking level-trigger events until they can
> >> be handled again, it would have to know which events are needed by
> >> libubus for internal processing of the pending request, etc.
> >
> > I though about adding something like a block_all_events_except_bla
> > to libubox and use it in libubus while waiting for the ubus method status
> > (if called with a special flag or so).
> I thought about this some more, and I think i'll rework libubus to no
> longer call uloop_run() for blocking calls. I'll make it use poll() instead.
>

I think this might also be useful for the lua mapping, because I think
what happens
in there isn't always 100% correct. ie. if you register a callback
from lua and then
issue a ubus call from the callback function, it's possible that the
ubus call will
generate a new callback before the current lua function has finished.
I'm not sure,
but I think this isn't allowed in the lua vm.

It was also the source of a previous bug, which was fixed by a patch I
send before:

[OpenWrt-Devel] [PATCH] ubus: Fix imbalance in lua stack push/pop of values.

https://lists.openwrt.org/pipermail/openwrt-devel/2014-February/023790.html
http://nbd.name/gitweb.cgi?p=luci2/ubus.git;a=commit;h=334c38918063b5ece164624e20490021b2bff38a

Karl
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] [PATCH] Allow installation of symlinks using RAMFS_COPY_BIN env var.

2014-03-31 Thread Karl Vogel 
run_ramfs() uses install_bin to copy executable files to the ram
partition. The install_bin function takes multiple parameters,
the first being the executable to copy and the rest of the args
are symlinks to that binary.

There currently isn't any way to pass multiple arguments to
install_bin in run_ramfs using the RAMFS_COPY_BIN parameter.

This patch enables specifying multiple parameters by separating
them with a colon. In effect one can now install a binary and it's
symlinks using RAMFS_COPY_BIN.

For example:
 RAMFS_COPY_BIN="/bin/busybox:/bin/sed:/bin/mktemp"

Will install 2 symlinks for sed and mktemp to the busybox binary
in the ramfs.

Without this patch, the sed and mktemp would be complete copies
of busybox, taking up precious ram memory.
---
 package/base-files/files/lib/upgrade/common.sh |2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/package/base-files/files/lib/upgrade/common.sh 
b/package/base-files/files/lib/upgrade/common.sh
index e53e844..31857a8 100644
--- a/package/base-files/files/lib/upgrade/common.sh
+++ b/package/base-files/files/lib/upgrade/common.sh
@@ -60,7 +60,7 @@ run_ramfs() { #  [...]
install_bin /sbin/fs-state
install_bin /sbin/snapshot
for file in $RAMFS_COPY_BIN; do
-   install_bin $file
+   install_bin ${file//:/ }
done
install_file /etc/resolv.conf /lib/functions.sh /lib/functions.sh 
/lib/upgrade/*.sh $RAMFS_COPY_DATA
 
-- 
1.7.1
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

[OpenWrt-Devel] CONFIG_DEBUG and CONFIG_USE_SSTRIP

2014-06-17 Thread Karl Vogel 
Currently when CONFIG_DEBUG is enabled, stripping is
disabled and the sstrip option becomes unavailable.

Looking at the commit history I noticed ticket #6373 which
introduced some of these changes. Reading the ticket it
seems the intend was to disable stripping when a user
selected CONFIG_DEBUG.

However the combination of debug + stripping is useful to
create a minimal build but still have the option to remote
gdb or examine core dumps, ie where the target has
stripped binaries, but the unstripped binaries are available
on the host/dev system. So I was wondering if it was
intentional to remove the sstrip option (as GNU strip is still
available) to work around bugs or issues with this
combination? or was it an unfortunate side effect of fixing
this ticket?

Karl
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel