Skip to contentSkip to conversation

x264: needs mmu

PATCH New Gustavo Zacarias 2 participants 6 messages May 2, 2016

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x264: needs mmu

Uses madvise() which is not available in NOMMU. Fixes:
http://autobuild.buildroot.net/results/837/837fd5a63d59b5c65818ec005a565cb7741a1cdd/

---
 package/x264/Config.in | 1 +
 1 file changed, 1 insertion(+)
package/x264/Config.in +1 -0
index be3ab3f..d042ddc 100644
@@ -1,5 +1,6 @@
 config BR2_PACKAGE_X264
 	bool "x264"
+	depends on BR2_USE_MMU # madvise()
 	help
 	  x264 is a free software library and application for
 	  encoding video streams into the H.264/MPEG-4 AVC
Signature · 1 line
2.7.3
#1 lore raw

PATCH 1/1 x264: needs mmu

Quoted 4 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Gustavo" == Gustavo Zacarias <[email protected]> writes:

Uses madvise() which is not available in NOMMU. Fixes:
http://autobuild.buildroot.net/results/837/837fd5a63d59b5c65818ec005a565cb7741a1cdd/

Hmm, is that really generally true? I don't quite see why you cannot have madvise() without mmu. Looking at uClibc-ng, the option enabling madvise support doesn't depend on MMU support.

Signature · 1 line
Bye, Peter Korsgaard
#2 lore raw

PATCH 1/1 x264: needs mmu

Quoted 3 lines · Peter Korsgaard

Hmm, is that really generally true? I don't quite see why you cannot have madvise() without mmu. Looking at uClibc-ng, the option enabling madvise support doesn't depend on MMU support.

I've based the rule on previous 80b9dfce, 5b309e1a and 8d446718 scenarios and didn't investigate myself.
Regards.

#3 lore raw

PATCH 1/1 x264: needs mmu

Quoted 9 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Gustavo" == Gustavo Zacarias <[email protected]> writes:

Quoted 3 lines · Peter Korsgaard

Hmm, is that really generally true? I don't quite see why you cannot have madvise() without mmu. Looking at uClibc-ng, the option enabling madvise support doesn't depend on MMU support.

I've based the rule on previous 80b9dfce, 5b309e1a and 8d446718 scenarios and didn't investigate myself.

Ok - Thomas, do you know if madvise generally isn't available or if it is something specific to the binary bfin toolchain?

I'll kick off a build with our new stm32f defconfig and look.

Signature · 2 lines
Venlig hilsen,
Peter Korsgaard 
#4 lore raw

PATCH 1/1 x264: needs mmu

Quoted 4 lines · Peter Korsgaard

Ok - Thomas, do you know if madvise generally isn't available or if it is something specific to the binary bfin toolchain?

I'll kick off a build with our new stm32f defconfig and look.

Ostensibly madvise() can be thrown away if the target doesn't support it, though it would require submitting a bunch of patches upstream ;) Regards.

#5 lore raw

PATCH 1/1 x264: needs mmu

In reply to Peter Korsgaard
Quoted 8 lines
Quoted 7 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

"Peter" == Peter Korsgaard <[email protected]> writes:

"Gustavo" == Gustavo Zacarias <[email protected]> writes:

Quoted 3 lines · Peter Korsgaard

Hmm, is that really generally true? I don't quite see why you cannot have madvise() without mmu. Looking at uClibc-ng, the option enabling madvise support doesn't depend on MMU support.

Quoted 7 lines
Quoted 2 lines

I've based the rule on previous 80b9dfce, 5b309e1a and 8d446718 scenarios and didn't investigate myself.

Ok - Thomas, do you know if madvise generally isn't available or if it is something specific to the binary bfin toolchain?

I'll kick off a build with our new stm32f defconfig and look.

As expected, building stm32f429_disco_defconfig with x264 works:

arm-linux-nm staging/usr/lib/libc.a | egrep 'fork|madvise' vfork.o:
00000000 T __GI_vfork
00000000 W vfork
00000000 T __vfork
madvise.o:
00000000 T madvise
posix_madvise.o:
00000000 T posix_madvise
U __GI_vfork
00000000 T _fork_parent
U __GI_vfork

So I'll change it to simply disable for that bfin toolchain instead.

Signature · 1 line
Bye, Peter Korsgaard