Skip to contentSkip to conversation

2.5.56: undefined reference to `_ebss' from drivers/mtd/maps/uclinux.c

DISCUSSION Adrian Bunk 2 participants 5 messages Jan 12, 2003 – Mar 21, 2003 Memory technology devices (MTD)

lore
Share & tools
lore
#0 lore raw

Hi Greg,

trying to compile 2.5.56 with CONFIG_MTD_UCLINUX fails on i386 with
undefined reference to `_ebss'
at the final linking.

It seems _ebss is only defined on the architectures m68knommu and v850?

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#1 lore raw

Hi Adrian,

Quoted 5 lines · Adrian Bunk

trying to compile 2.5.56 with CONFIG_MTD_UCLINUX fails on i386 with
undefined reference to `_ebss'
at the final linking.

It seems _ebss is only defined on the architectures m68knommu and v850?

Hmm, currently that is correct. There doesn't appear to be a "standard" symbol name applied to the immediate end of the bss section. Different architectures are using different names:

  _ebss        -- m68knommu, v850
  __bss_stop   -- i386, alpha, ppc, s390
  __bss_end    -- x86_64
  _end         -- mips, parisc, sparc, (actually most have this)

Actually it looks like _end is probably closer, it seems to almost always fall strait after the bss, on just about every architecture that has it.

Come to think of it _end is probably more appropriate anyway. Since that code is trying to find the location of something concatenated to the end of the kernel image.

Regards
Greg

------------------------------------------------------------------------
Greg Ungerer -- Chief Software Wizard EMAIL: [email protected]
SnapGear Pty Ltd PHONE: +61 7 3435 2888
825 Stanley St, FAX: +61 7 3891 3630
Woolloongabba, QLD, 4102, Australia WEB: www.SnapGear.com

#2 lore raw
Quoted 18 lines · Greg Ungerer

Hi Adrian,

Quoted 5 lines · Adrian Bunk

trying to compile 2.5.56 with CONFIG_MTD_UCLINUX fails on i386 with
undefined reference to `_ebss'
at the final linking.

It seems _ebss is only defined on the architectures m68knommu and v850?

Hmm, currently that is correct. There doesn't appear to be a "standard" symbol name applied to the immediate end of the bss section. Different architectures are using different names:

  _ebss        -- m68knommu, v850
  __bss_stop   -- i386, alpha, ppc, s390
  __bss_end    -- x86_64
  _end         -- mips, parisc, sparc, (actually most have this)
Quoted 7 lines

Actually it looks like _end is probably closer, it seems to almost always fall strait after the bss, on just about every architecture that has it.

Come to think of it _end is probably more appropriate anyway. Since that code is trying to find the location of something concatenated to the end of the kernel image.

This problem (undefined reference to `_ebss') is still present in 2.5.65.

It might not be a solution for the whole issue, but is it intentionally that it's possible to enable CONFIG_MTD_UCLINUX on non-uClinux architectures or should an appropriate dependency be added to the Kconfig file?

Quoted 2 lines

Regards
Greg

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#3 lore raw
Quoted 5 lines · Adrian Bunk

...
It might not be a solution for the whole issue, but is it intentionally that it's possible to enable CONFIG_MTD_UCLINUX on non-uClinux architectures or should an appropriate dependency be added to the Kconfig file?

The following patch should do what I was thinking of:

drivers/mtd/maps/Kconfig renamed from drivers/mtd/maps/Kconfig.old +1 -1
@@ -348,7 +348,7 @@
 
 config MTD_UCLINUX
 	tristate "Generic uClinux RAM/ROM filesystem support"
-	depends on MTD_PARTITIONS
+	depends on MTD_PARTITIONS && !MMU
 	help
 	  Map driver to support image based filesystems for uClinux.
 

Any comments on whether this is correct?

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#4 lore raw

Hi Adrian,

This seems reasonable to me. It is after all targeted at uclinux. I suspect noone with VM setups would be much interrested in using it anyway.

Patch looks good to me.

Regards
Greg

Quoted 28 lines · Adrian Bunk
Quoted 3 lines · Adrian Bunk

...
It might not be a solution for the whole issue, but is it intentionally
that it's possible to enable CONFIG_MTD_UCLINUX on non-uClinux architectures or should an appropriate dependency be added to the Kconfig file?

The following patch should do what I was thinking of:

drivers/mtd/maps/Kconfig renamed from drivers/mtd/maps/Kconfig.old +1 -1
@@ -348,7 +348,7 @@
 
 config MTD_UCLINUX
 	tristate "Generic uClinux RAM/ROM filesystem support"
-	depends on MTD_PARTITIONS
+	depends on MTD_PARTITIONS && !MMU
 	help
 	  Map driver to support image based filesystems for uClinux.
 

Any comments on whether this is correct?

cu
Adrian

Signature · 5 lines
------------------------------------------------------------------------
Greg Ungerer  --  Chief Software Wizard        EMAIL:  [email protected]
SnapGear Pty Ltd                               PHONE:    +61 7 3435 2888
825 Stanley St,                                  FAX:    +61 7 3891 3630
Woolloongabba, QLD, 4102, Australia              WEB:   www.SnapGear.com