firmware issues #2?

DISCUSSION Applied Thomas Penteker 3 participants 3 messages Apr 23, 2010 – Apr 24, 2010

lore

Share & tools

lore

Thomas Penteker

author

2010-04-23 16:55 UTC #0 lore raw

Hello,

calling configure of udev-153 gives me the following:

% ./configure --disable-extras --disable-introspection --with-firmware-path=/lib/SOMETHINGUNIQUE:/lib/foo (...)
firmware path:
(...)

Compliation proves this:

% make V=1 2>&1 | grep -- -DFIRMWARE
gcc -DHAVE_CONFIG_H -I. -include ./config.h -I./libudev -DSYSCONFDIR=\""/usr/etc"\" -DLIBEXECDIR=\""/usr/libexec"\" -DFIRMWARE_PATH="" -g -O2 -MT extras/firmware/extras_firmware_firmware-firmware.o -MD -MP -MF extras/firmware/.deps/extras_firmware_firmware-firmware.Tpo -c -o extras/firmware/extras_firmware_firmware-firmware.o test -f 'extras/firmware/firmware.c' || echo './'extras/firmware/firmware.c

Correcting the FIRMWARE_PATH variable in Makefile fixes the issue.

So I think the firmware-path still is broken (but in a different way).

kind regards, Thomas Penteker

Signature · 1 line

dynamic software linking table corrupted

Robby Workman

2010-04-23 19:25 UTC #1 lore raw

Quoted 20 lines · On Fri, 23 Apr 2010 18:55:24 +0200 Thomas Penteker <[email protected]> wrote:

Hello,
calling configure of udev-153 gives me the following:
% ./configure --disable-extras --disable-introspection --with-firmware-path=/lib/SOMETHINGUNIQUE:/lib/foo (...)
firmware path:
(...)
Compliation proves this:
% make V=1 2>&1 | grep -- -DFIRMWARE
gcc -DHAVE_CONFIG_H -I. -include ./config.h -I./libudev
(quoted diff) +0 -2
-DSYSCONFDIR=\""/usr/etc"\" -DLIBEXECDIR=\""/usr/libexec"\"
-DFIRMWARE_PATH="" -g -O2 -MT
extras/firmware/extras_firmware_firmware-firmware.o -MD -MP -MF extras/firmware/.deps/extras_firmware_firmware-firmware.Tpo -c -o extras/firmware/extras_firmware_firmware-firmware.o test -f 'extras/firmware/firmware.c' || echo './'extras/firmware/firmware.c

Quoted 3 lines

Correcting the FIRMWARE_PATH variable in Makefile fixes the issue.
So I think the firmware-path still is broken (but in a different way).

I'm not sure it's the *correct* fix, but the attached patch seems to fix it here.

http://git.rlworkman.net/?p=udev.git;a=commit;h=2d85e63bfb028f7a4010ace1c1e7da1220d557cb

-RW

commit 2d85e63bfb028f7a4010ace1c1e7da1220d557cb

Author: Robby Workman <[email protected]>
Date:   Fri Apr 23 14:19:23 2010 -0500

configure.ac: move firmware path setting out of extras nest

configure.ac +20 ~~-20~~

index d4d29c4..ccf6221 100644
@@ -12,6 +12,26 @@ GTK_DOC_CHECK(1.10)
 AC_PREFIX_DEFAULT([/usr])
 AC_PATH_PROG([XSLTPROC], [xsltproc])
 
+AC_ARG_WITH(firmware-path,
+	AS_HELP_STRING([--with-firmware-path=DIR[[[:DIR[...]]]]],
+	   [Firmware search path (default=/lib/firmware/updates:/lib/firmware)]),
+	[],
+	[with_firmware_path="/lib/firmware/updates:/lib/firmware"]
+	)
+OLD_IFS=$IFS
+IFS=:
+for i in $with_firmware_path
+do
+	if test "x${FIRMWARE_PATH}" = "x"
+	then
+		FIRMWARE_PATH="\\\"${i}/\\\""
+	else
+		FIRMWARE_PATH="${FIRMWARE_PATH}, \\\"${i}/\\\""
+	fi
+done
+IFS=$OLD_IFS
+AC_SUBST([FIRMWARE_PATH], [$FIRMWARE_PATH])
+
 AC_ARG_WITH([rootlibdir],
 	AS_HELP_STRING([--with-rootlibdir=DIR], [rootfs directory to install shared libraries]),
 	[], [with_rootlibdir=$libdir])

@@ -82,26 +102,6 @@ if test "x$enable_extras" = xyes; then
 		fi])
 	AC_SUBST(PCI_DATABASE)
 
-	AC_ARG_WITH(firmware-path,
-		AS_HELP_STRING([--with-firmware-path=DIR[[[:DIR[...]]]]],
-		   [Firmware search path (default=/lib/firmware/updates:/lib/firmware)]),
-		[],
-		[with_firmware_path="/lib/firmware/updates:/lib/firmware"]
-		)
-	OLD_IFS=$IFS
-	IFS=:
-	for i in $with_firmware_path
-	do
-		if test "x${FIRMWARE_PATH}" = "x"
-		then
-			FIRMWARE_PATH="\\\"${i}/\\\""
-		else
-			FIRMWARE_PATH="${FIRMWARE_PATH}, \\\"${i}/\\\""
-		fi
-	done
-	IFS=$OLD_IFS
-	AC_SUBST([FIRMWARE_PATH], [$FIRMWARE_PATH])
-
 	AC_CHECK_HEADER([linux/input.h], [:], AC_MSG_ERROR([kernel headers not found]))
 	AC_SUBST([INCLUDE_PREFIX], [$(echo '#include <linux/input.h>' | eval $ac_cpp -E - | sed -n '/linux\/input.h/ {s:.*"\(.*\)/linux/input.h".*:\1:; p; q}')])
 fi

Kay Sievers

Applied

2010-04-24 19:42 UTC #2 lore raw

In reply to Thomas Penteker

Quoted 19 lines · Robby Workman

Quoted 18 lines · On Fri, 23 Apr 2010 18:55:24 +0200 Thomas Penteker <[email protected]> wrote:
calling configure of udev-153 gives me the following:
% ./configure --disable-extras --disable-introspection --with-firmware-path=/lib/SOMETHINGUNIQUE:/lib/foo (...) firmware path:
(...)
Compliation proves this:
% make V=1 2>&1 | grep -- -DFIRMWARE
gcc -DHAVE_CONFIG_H -I. -include ./config.h -I./libudev
(quoted diff) +0 -2
-DSYSCONFDIR=\""/usr/etc"\" -DLIBEXECDIR=\""/usr/libexec"\"
-DFIRMWARE_PATH="" -g -O2 -MT
extras/firmware/extras_firmware_firmware-firmware.o -MD -MP -MF extras/firmware/.deps/extras_firmware_firmware-firmware.Tpo -c -o extras/firmware/extras_firmware_firmware-firmware.o test -f 'extras/firmware/firmware.c' || echo './'extras/firmware/firmware.c

Quoted 8 lines

Quoted 3 lines
Correcting the FIRMWARE_PATH variable in Makefile fixes the issue.
So I think the firmware-path still is broken (but in a different way).
I'm not sure it's the *correct* fix, but the attached patch seems to fix it here.
http://git.rlworkman.net/?p=udev.git;a=commit;h-85e63bfb028f7a4010ace1c1e7da1220d557cb

Applied.

Thanks,
Kay