Patchwork [U-Boot] Makefile: prevent libgcc to be linked twice

login
register
mail settings
Submitter Nicolas Ferre
Date Oct. 17, 2011, 1:41 p.m.
Message ID <1318858875-14201-1-git-send-email-nicolas.ferre@atmel.com>
Download mbox | patch
Permalink /patch/120213/
State Rejected
Delegated to: Mike Frysinger
Headers show

Comments

Nicolas Ferre - Oct. 17, 2011, 1:41 p.m.
If we define USE_PRIVATE_LIBGCC, PLATFORM_LIBS is used during
link. During this last link editing, libgcc may be listed twice
and fail.
Prevent this using "filter-out" in top Makefile.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
---
 Makefile |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)
Mike Frysinger - Oct. 17, 2011, 5:25 p.m.
On Monday 17 October 2011 09:41:15 Nicolas Ferre wrote:
> If we define USE_PRIVATE_LIBGCC, PLATFORM_LIBS is used during
> link. During this last link editing, libgcc may be listed twice
> and fail.
> Prevent this using "filter-out" in top Makefile.

could you please provide more details ?  libgcc should not be listed twice.
-mike
Nicolas Ferre - Oct. 18, 2011, 8:30 a.m.
On 10/17/2011 07:25 PM, Mike Frysinger :
> On Monday 17 October 2011 09:41:15 Nicolas Ferre wrote:
>> If we define USE_PRIVATE_LIBGCC, PLATFORM_LIBS is used during
>> link. During this last link editing, libgcc may be listed twice
>> and fail.
>> Prevent this using "filter-out" in top Makefile.
> 
> could you please provide more details ?  libgcc should not be listed twice.

I compile current u-boot with the flowing command line:
make -j5 CROSS_COMPILE=arm-linux-gnueabi- USE_PRIVATE_LIBGCC=yes at91sam9m10g45ek_nandflash 

On an Ubuntu 11.10 with Linaro's cross-toolchain (targeting ARMv7). My SoC is an ARM926 so I use the "USE_PRIVATE_LIBGCC=yes" switch. This will prevent the build to take libgcc from my host but compile the u-boot embedded libgcc: everything is fine, good feature.

But, here is the final link editing stage:

UNDEF_SYM=`arm-linux-gnueabi-objdump -x board/atmel/at91sam9m10g45ek/libat91sam9m10g45ek.o api/libapi.o arch/arm/cpu/arm926ejs/at91/libat91.o arch/arm/cpu/arm926ejs/libarm926ejs.o arch/arm/lib/libarm.o common/libcommon.o disk/libdisk.o drivers/bios_emulator/libatibiosemu.o drivers/block/libblock.o drivers/dma/libdma.o drivers/fpga/libfpga.o drivers/gpio/libgpio.o drivers/hwmon/libhwmon.o drivers/i2c/libi2c.o drivers/input/libinput.o drivers/misc/libmisc.o drivers/mmc/libmmc.o drivers/mtd/libmtd.o drivers/mtd/nand/libnand.o drivers/mtd/onenand/libonenand.o drivers/mtd/spi/libspi_flash.o drivers/mtd/ubi/libubi.o drivers/net/libnet.o drivers/net/phy/libphy.o drivers/pci/libpci.o drivers/pcmcia/libpcmcia.o drivers/power/libpower.o drivers/rtc/librtc.o drivers/serial/libserial.o drivers/spi/libspi.o drivers/twserial/libtws.o drivers/usb/eth/libusb_eth.o drivers/usb/gadget/libusb_gadget.o drivers/usb/host/libusb_host.o drivers/usb/musb/libusb_musb.o drivers/usb/phy/libusb_phy.o dri
vers/video/libvideo.o drivers/watchdog/libwatchdog.o fs/cramfs/libcramfs.o fs/ext2/libext2fs.o fs/fat/libfat.o fs/fdos/libfdos.o fs/jffs2/libjffs2.o fs/reiserfs/libreiserfs.o fs/ubifs/libubifs.o fs/yaffs2/libyaffs2.o lib/libfdt/libfdt.o lib/libgeneric.o lib/lzma/liblzma.o lib/lzo/liblzo.o lib/zlib/libz.o net/libnet.o post/libpost.o | sed  -n -e 's/.*\(__u_boot_cmd_.*\)/-u\1/p'|sort|uniq`; cd /home/nferre/workspace/u-boot/u-boot-git && arm-linux-gnueabi-ld  -pie -T u-boot.lds -Bstatic -Ttext 0x73f00000 $UNDEF_SYM arch/arm/cpu/arm926ejs/start.o --start-group api/libapi.o arch/arm/cpu/arm926ejs/at91/libat91.o arch/arm/cpu/arm926ejs/libarm926ejs.o arch/arm/lib/libarm.o common/libcommon.o disk/libdisk.o drivers/bios_emulator/libatibiosemu.o drivers/block/libblock.o drivers/dma/libdma.o drivers/fpga/libfpga.o drivers/gpio/libgpio.o drivers/hwmon/libhwmon.o drivers/i2c/libi2c.o drivers/input/libinput.o drivers/misc/libmisc.o drivers/mmc/libmmc.o drivers/mtd/libmtd.o drivers/mtd/nand
/libnand.o drivers/mtd/onenand/libonenand.o drivers/mtd/spi/libspi_flash.o drivers/mtd/ubi/libubi.o drivers/net/libnet.o drivers/net/phy/libphy.o drivers/pci/libpci.o drivers/pcmcia/libpcmcia.o drivers/power/libpower.o drivers/rtc/librtc.o drivers/serial/libserial.o drivers/spi/libspi.o drivers/twserial/libtws.o drivers/usb/eth/libusb_eth.o drivers/usb/gadget/libusb_gadget.o drivers/usb/host/libusb_host.o drivers/usb/musb/libusb_musb.o drivers/usb/phy/libusb_phy.o drivers/video/libvideo.o drivers/watchdog/libwatchdog.o fs/cramfs/libcramfs.o fs/ext2/libext2fs.o fs/fat/libfat.o fs/fdos/libfdos.o fs/jffs2/libjffs2.o fs/reiserfs/libreiserfs.o fs/ubifs/libubifs.o fs/yaffs2/libyaffs2.o lib/libfdt/libfdt.o lib/libgeneric.o lib/lzma/liblzma.o lib/lzo/liblzo.o lib/zlib/libz.o net/libnet.o post/libpost.o board/atmel/at91sam9m10g45ek/libat91sam9m10g45ek.o --end-group /home/nferre/workspace/u-boot/u-boot-git/arch/arm/lib/eabi_compat.o /home/nferre/workspace/u-boot/u-boot-git/arch/arm/lib
/libgcc.o /home/nferre/workspace/u-boot/u-boot-git/arch/arm/lib/libgcc.o -Map u-boot.map -o u-boot


And you see that there is arch/arm/lib/libgcc.o included twice in object files that have to be linked together. I think that it is coming from the PLATFORM_LIBS variable so I modified like in my patch which solves the problem.

Best regards,
Mike Frysinger - Oct. 18, 2011, 4:06 p.m.
On Tuesday 18 October 2011 04:30:27 Nicolas Ferre wrote:
> On 10/17/2011 07:25 PM, Mike Frysinger :
> > On Monday 17 October 2011 09:41:15 Nicolas Ferre wrote:
> >> If we define USE_PRIVATE_LIBGCC, PLATFORM_LIBS is used during
> >> link. During this last link editing, libgcc may be listed twice
> >> and fail.
> >> Prevent this using "filter-out" in top Makefile.
> > 
> > could you please provide more details ?  libgcc should not be listed
> > twice.
> 
> I compile current u-boot with the flowing command line:
> make -j5 CROSS_COMPILE=arm-linux-gnueabi- USE_PRIVATE_LIBGCC=yes
> at91sam9m10g45ek_nandflash

the issue shows up when using the board shortcut, and only when your tree is 
already configured.  if you use at91sam9m10g45ek_nandflash_config and then 
another make, it works fine.  if you start from a clean tree (git clean -x -d), 
it works fine.

this is due to PLATFORM_LIBS appending PLATFORM_LIBGCC when the tree has been 
configured, then exporting PLATFORM_LIBS, then the board shortcut running 
$(MAKE).  the exported PLATFORM_LIBS gets LIBGCC appended a second time.

this issue will show up with any value that gets appended to PLATFORM_LIBS and 
not just libgcc.  the arm code already hits this and deals with it locally.  
seems like we should instead unify this logic, or redo PLATFORM_LIBS so that 
it isn't an issue in the first place.
-mike
Nicolas Ferre - Oct. 20, 2011, 8:26 a.m.
On 10/18/2011 06:06 PM, Mike Frysinger :
> On Tuesday 18 October 2011 04:30:27 Nicolas Ferre wrote:
>> On 10/17/2011 07:25 PM, Mike Frysinger :
>>> On Monday 17 October 2011 09:41:15 Nicolas Ferre wrote:
>>>> If we define USE_PRIVATE_LIBGCC, PLATFORM_LIBS is used during
>>>> link. During this last link editing, libgcc may be listed twice
>>>> and fail.
>>>> Prevent this using "filter-out" in top Makefile.
>>>
>>> could you please provide more details ?  libgcc should not be listed
>>> twice.
>>
>> I compile current u-boot with the flowing command line:
>> make -j5 CROSS_COMPILE=arm-linux-gnueabi- USE_PRIVATE_LIBGCC=yes
>> at91sam9m10g45ek_nandflash
> 
> the issue shows up when using the board shortcut, and only when your tree is 
> already configured.  if you use at91sam9m10g45ek_nandflash_config and then 
> another make, it works fine.  if you start from a clean tree (git clean -x -d), 
> it works fine.

Ah thanks for the information. Indeed it is a more common way of using
the configuration/compilation flow...

> this is due to PLATFORM_LIBS appending PLATFORM_LIBGCC when the tree has been 
> configured, then exporting PLATFORM_LIBS, then the board shortcut running 
> $(MAKE).  the exported PLATFORM_LIBS gets LIBGCC appended a second time.
> 
> this issue will show up with any value that gets appended to PLATFORM_LIBS and 
> not just libgcc.  the arm code already hits this and deals with it locally.  
> seems like we should instead unify this logic, or redo PLATFORM_LIBS so that 
> it isn't an issue in the first place.

Yes, that it a bit confusing... But with proper explanation, we can deal
with this.

Thanks for your help, best regards,

Patch

diff --git a/Makefile b/Makefile
index 5db2e0e..4bdfe11 100644
--- a/Makefile
+++ b/Makefile
@@ -311,7 +311,8 @@  endif
 else
 PLATFORM_LIBGCC = -L $(shell dirname `$(CC) $(CFLAGS) -print-libgcc-file-name`) -lgcc
 endif
-PLATFORM_LIBS += $(PLATFORM_LIBGCC)
+PLATFORM_LIBS := $(PLATFORM_LIBGCC) \
+	$(filter-out %$(PLATFORM_LIBGCC), $(PLATFORM_LIBS))
 export PLATFORM_LIBS
 
 # Special flags for CPP when processing the linker script.