Skip to contentSkip to conversation

ARM Kernel autobuilder

PATCH New Vincent Sanders 7 participants 16 messages Jul 22, 2010 – Aug 3, 2010

Subsystems (2)
Arm (32-bit) Arm sub-architectures
lore
Share & tools
lore

Vincent Sanders

author
#0 lore raw

A short note to apologize to the kirkwood people over the autobuilder producing junk results for their defconfig on the last few builds. The ccache got corrupted due to a disk glitch. All fixed now.

The current level of 177 Configurations is (to put it mildly) insane. The 5 year report posted previously explained the specifics. suffice to say with Linus recent activities anything we can do to reduce this count is useful.

Can I propose now the omap3_defconfig is working that the duplicate defconfigs are removed, these are:

cm_t35_defconfig
devkit8000_defconfig
igep0020_defconfig
rx51_defconfig
omap3_beagle_defconfig
omap3_evm_defconfig
am3517_evm_defconfig
omap_4430sdp_defconfig
omap_ldp_defconfig
omap3_pandora_defconfig
omap3_stalker_lks_defconfig
omap_apollon_2420_defconfig
omap_h4_2420_defconfig
omap_2430sdp_defconfig
omap_3430sdp_defconfig
omap_3630sdp_defconfig
omap3_touchbook_defconfig
omap_zoom2_defconfig
omap_zoom3_defconfig
overo_defconfig

This would reduce the default configuration count to 156 configurations.

There are still a huge number of configurations which support a single machine which I may look at creating additional combined targets for futher reducing the build count.

Finally the n8x0 and ns9xxx configurations have failed to build for a while now.

The n8x0 config supports machines maintained by Kai Svahn, Lauri Leukkunen and Timo O. Karjalainen can one of you please fix the build?

The ns9xxx configuration has been failing for even longer and teh machines supported are maintainerd by Uwe Kleine-Konig

Signature · 9 lines
Regards Vincent
http://www.kyllikki.org/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20100722/6bdebb07/attachment.sig>
#1 lore raw

Vincent Sanders <[email protected]> writes:

Hi,

Quoted 3 lines

A short note to apologize to the kirkwood people over the autobuilder producing junk results for their defconfig on the last few builds. The ccache got corrupted due to a disk glitch. All fixed now.

Thanks a lot for fixing that (I was really wondering what was the issue as I couldn't reproduce it here) and thanks even more for all the work you've done on the kautobuild stuff.

[...]

Quoted 5 lines

Finally the n8x0 and ns9xxx configurations have failed to build for a while now.

The n8x0 config supports machines maintained by Kai Svahn, Lauri Leukkunen and Timo O. Karjalainen can one of you please fix the build?

It has been discussed on the linux-omap mailing list :
http://marc.info/?l=linux-omap&m=127832134917694&w=2

Don't know the status of the patch though.

Arnaud

#2 lore raw
Quoted 5 lines · Arnaud Patard
Quoted 3 lines · Vincent Sanders

Finally the n8x0 and ns9xxx configurations have failed to build for a while now.

Quoted 7 lines
Quoted 2 lines

The n8x0 config supports machines maintained by Kai Svahn, Lauri Leukkunen and Timo O. Karjalainen can one of you please fix the build?

It has been discussed on the linux-omap mailing list :
http://marc.info/?l=linux-omap&m=127832134917694&w=2

Don't know the status of the patch though.

That's merged as 2b795ea00c2bbb077a1199a4d729c8ac03a6bded.

Regards,

Tony

#3 lore raw
In reply to Vincent Sanders
Quoted 24 lines · Vincent Sanders

Can I propose now the omap3_defconfig is working that the duplicate defconfigs are removed, these are:

cm_t35_defconfig
devkit8000_defconfig
igep0020_defconfig
rx51_defconfig
omap3_beagle_defconfig
omap3_evm_defconfig
am3517_evm_defconfig
omap_4430sdp_defconfig
omap_ldp_defconfig
omap3_pandora_defconfig
omap3_stalker_lks_defconfig
omap_apollon_2420_defconfig
omap_h4_2420_defconfig
omap_2430sdp_defconfig
omap_3430sdp_defconfig
omap_3630sdp_defconfig
omap3_touchbook_defconfig
omap_zoom2_defconfig
omap_zoom3_defconfig
overo_defconfig

Quoted 2 lines

This would reduce the default configuration count to 156 configurations.

Linus lately appears to be more relaxed about this, and is in the mood for improvements in the situation, so yes this would be an improvement. It certainly would make things a little smaller, and allow kautobuild to run faster.

#4 lore raw
Quoted 33 lines

-----Original Message-----
From: linux-arm-kernel-bounces at lists.infradead.org [mailto:linux-arm- kernel-bounces at lists.infradead.org] On Behalf Of Russell King - ARM Linux
Sent: Thursday, July 22, 2010 6:54 PM
To: Vincent Sanders
Cc: linux-arm-kernel at lists.infradead.org
Subject: Re: ARM Kernel autobuilder

Quoted 24 lines · Vincent Sanders

Can I propose now the omap3_defconfig is working that the duplicate defconfigs are removed, these are:

cm_t35_defconfig
devkit8000_defconfig
igep0020_defconfig
rx51_defconfig
omap3_beagle_defconfig
omap3_evm_defconfig
am3517_evm_defconfig
omap_4430sdp_defconfig
omap_ldp_defconfig
omap3_pandora_defconfig
omap3_stalker_lks_defconfig
omap_apollon_2420_defconfig
omap_h4_2420_defconfig
omap_2430sdp_defconfig
omap_3430sdp_defconfig
omap_3630sdp_defconfig
omap3_touchbook_defconfig
omap_zoom2_defconfig
omap_zoom3_defconfig
overo_defconfig

Quoted 8 lines
Quoted 2 lines

This would reduce the default configuration count to 156 configurations.

Linus lately appears to be more relaxed about this, and is in the mood for improvements in the situation, so yes this would be an improvement. It certainly would make things a little smaller, and allow kautobuild to run faster.

omap_4430sdp_defconfig is not duplicate. Please keep omap_4430sdp_defconfig because omap3_defconfig doen't includes SMP build.

Regards,
Santosh

#5 lore raw
Quoted 29 lines · Santosh
Quoted 28 lines · Russell King - ARM Linux
Quoted 26 lines · Vincent Sanders

Can I propose now the omap3_defconfig is working that the duplicate defconfigs are removed, these are:

cm_t35_defconfig
devkit8000_defconfig
igep0020_defconfig
rx51_defconfig
omap3_beagle_defconfig
omap3_evm_defconfig
am3517_evm_defconfig
omap_4430sdp_defconfig
omap_ldp_defconfig
omap3_pandora_defconfig
omap3_stalker_lks_defconfig
omap_apollon_2420_defconfig
omap_h4_2420_defconfig
omap_2430sdp_defconfig
omap_3430sdp_defconfig
omap_3630sdp_defconfig
omap3_touchbook_defconfig
omap_zoom2_defconfig
omap_zoom3_defconfig
overo_defconfig

This would reduce the default configuration count to 156 configurations.

Quoted 6 lines
Quoted 5 lines

Linus lately appears to be more relaxed about this, and is in the mood for improvements in the situation, so yes this would be an improvement. It certainly would make things a little smaller, and allow kautobuild to run faster.

omap_4430sdp_defconfig is not duplicate. Please keep omap_4430sdp_defconfig because omap3_defconfig doen't includes SMP build.

Are there any other reasons to keep omap_4430sdp_defconfig? SMP seems to be a weak argument to me, after all if I peek in to arch/x86/configs/ I don't see SMP and non-SMP variants.

#6 lore raw
Quoted 44 lines

-----Original Message-----
From: Simon Horman [mailto:horms at verge.net.au]
Sent: Friday, July 23, 2010 9:35 AM
To: Shilimkar, Santosh
Cc: Russell King - ARM Linux; Vincent Sanders; linux-arm- kernel at lists.infradead.org
Subject: Re: ARM Kernel autobuilder

Quoted 26 lines · Santosh
Quoted 25 lines · Russell King - ARM Linux
Quoted 24 lines · Vincent Sanders

Can I propose now the omap3_defconfig is working that the duplicate defconfigs are removed, these are:

cm_t35_defconfig
devkit8000_defconfig
igep0020_defconfig
rx51_defconfig
omap3_beagle_defconfig
omap3_evm_defconfig
am3517_evm_defconfig
omap_4430sdp_defconfig
omap_ldp_defconfig
omap3_pandora_defconfig
omap3_stalker_lks_defconfig
omap_apollon_2420_defconfig
omap_h4_2420_defconfig
omap_2430sdp_defconfig
omap_3430sdp_defconfig
omap_3630sdp_defconfig
omap3_touchbook_defconfig
omap_zoom2_defconfig
omap_zoom3_defconfig
overo_defconfig

Quoted 5 lines
Quoted 5 lines
Quoted 2 lines

This would reduce the default configuration count to 156 configurations.

Linus lately appears to be more relaxed about this, and is in the mood for improvements in the situation, so yes this would be an

improvement.

Quoted 3 lines
Quoted 3 lines

It certainly would make things a little smaller, and allow kautobuild to run faster.

Quoted 6 lines
Quoted 1 line

omap_4430sdp_defconfig is not duplicate. Please keep

omap_4430sdp_defconfig because omap3_defconfig doen't includes SMP build.

Are there any other reasons to keep omap_4430sdp_defconfig? SMP seems to be a weak argument to me, after all if I peek in to arch/x86/configs/ I don't see SMP and non-SMP variants.

OMAP4 is way different from OMAP3.

OMAP3:
Coretex-A8 UP
Internal L2 cache
DSP System
IVHA
OMAP4 -
Cortex-A9, Dual Core
External L2 cache ( PL310)
New Interrupt controller(GIC)
New Memory subsystem (EMIF)
3 Additional co-procesors
New DSS
IVHA2

Except ARMv7 architecture, there is not much common between these two SOCs

Regards,
Santosh

#7 lore raw
Quoted 53 lines

-----Original Message-----
From: linux-arm-kernel-bounces at lists.infradead.org [mailto:linux-arm- kernel-bounces at lists.infradead.org] On Behalf Of Shilimkar, Santosh
Sent: Friday, July 23, 2010 10:59 AM
To: Simon Horman
Cc: Vincent Sanders; Russell King - ARM Linux; linux-arm- kernel at lists.infradead.org
Subject: RE: ARM Kernel autobuilder

Quoted 8 lines

-----Original Message-----
From: Simon Horman [mailto:horms at verge.net.au]
Sent: Friday, July 23, 2010 9:35 AM
To: Shilimkar, Santosh
Cc: Russell King - ARM Linux; Vincent Sanders; linux-arm- kernel at lists.infradead.org
Subject: Re: ARM Kernel autobuilder

Quoted 4 lines
Quoted 3 lines · Santosh
Quoted 2 lines · Russell King - ARM Linux
Quoted 1 line · Vincent Sanders

Can I propose now the omap3_defconfig is working that the

duplicate

Quoted 23 lines
Quoted 23 lines
Quoted 23 lines
Quoted 23 lines

defconfigs are removed, these are:

cm_t35_defconfig
devkit8000_defconfig
igep0020_defconfig
rx51_defconfig
omap3_beagle_defconfig
omap3_evm_defconfig
am3517_evm_defconfig
omap_4430sdp_defconfig
omap_ldp_defconfig
omap3_pandora_defconfig
omap3_stalker_lks_defconfig
omap_apollon_2420_defconfig
omap_h4_2420_defconfig
omap_2430sdp_defconfig
omap_3430sdp_defconfig
omap_3630sdp_defconfig
omap3_touchbook_defconfig
omap_zoom2_defconfig
omap_zoom3_defconfig
overo_defconfig

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Quoted 2 lines

This would reduce the default configuration count to 156 configurations.

Linus lately appears to be more relaxed about this, and is in the

mood

Quoted 1 line
Quoted 1 line
Quoted 1 line

for improvements in the situation, so yes this would be an

Quoted 12 lines
Quoted 2 lines

improvement.

Quoted 1 line
Quoted 1 line

It certainly would make things a little smaller, and allow

kautobuild

Quoted 4 lines
Quoted 3 lines
Quoted 2 lines

to run faster.

omap_4430sdp_defconfig is not duplicate. Please keep

omap_4430sdp_defconfig because omap3_defconfig doen't includes SMP

build.

Quoted 4 lines

Are there any other reasons to keep omap_4430sdp_defconfig? SMP seems to be a weak argument to me, after all if I peek in to arch/x86/configs/ I don't see SMP and non-SMP variants.

x86 has only two defconfigs and both are SMP. So you won't find an UP machine there :) arch/x86/configs/i386_defconfig:CONFIG_SMP=y arch/x86/configs/x86_64_defconfig:CONFIG_SMP=y

Quoted 18 lines

OMAP4 is way different from OMAP3.

OMAP3:
Coretex-A8 UP
Internal L2 cache
DSP System
IVHA
OMAP4 -
Cortex-A9, Dual Core
External L2 cache ( PL310)
New Interrupt controller(GIC)
New Memory subsystem (EMIF)
3 Additional co-procesors
New DSS
IVHA2

Except ARMv7 architecture, there is not much common between these two SOCs

#8 lore raw
In reply to Shilimkar, Santosh
Quoted 66 lines · Santosh
Quoted 44 lines

-----Original Message-----
From: Simon Horman [mailto:horms at verge.net.au]
Sent: Friday, July 23, 2010 9:35 AM
To: Shilimkar, Santosh
Cc: Russell King - ARM Linux; Vincent Sanders; linux-arm- kernel at lists.infradead.org
Subject: Re: ARM Kernel autobuilder

Quoted 26 lines · Santosh
Quoted 25 lines · Russell King - ARM Linux
Quoted 24 lines · Vincent Sanders

Can I propose now the omap3_defconfig is working that the duplicate defconfigs are removed, these are:

cm_t35_defconfig
devkit8000_defconfig
igep0020_defconfig
rx51_defconfig
omap3_beagle_defconfig
omap3_evm_defconfig
am3517_evm_defconfig
omap_4430sdp_defconfig
omap_ldp_defconfig
omap3_pandora_defconfig
omap3_stalker_lks_defconfig
omap_apollon_2420_defconfig
omap_h4_2420_defconfig
omap_2430sdp_defconfig
omap_3430sdp_defconfig
omap_3630sdp_defconfig
omap3_touchbook_defconfig
omap_zoom2_defconfig
omap_zoom3_defconfig
overo_defconfig

Quoted 5 lines
Quoted 5 lines
Quoted 2 lines

This would reduce the default configuration count to 156 configurations.

Linus lately appears to be more relaxed about this, and is in the mood for improvements in the situation, so yes this would be an

improvement.

Quoted 3 lines
Quoted 3 lines

It certainly would make things a little smaller, and allow kautobuild to run faster.

Quoted 6 lines
Quoted 1 line

omap_4430sdp_defconfig is not duplicate. Please keep

omap_4430sdp_defconfig because omap3_defconfig doen't includes SMP build.

Are there any other reasons to keep omap_4430sdp_defconfig? SMP seems to be a weak argument to me, after all if I peek in to arch/x86/configs/ I don't see SMP and non-SMP variants.

OMAP4 is way different from OMAP3.

OMAP3:
Coretex-A8 UP
Internal L2 cache
DSP System
IVHA
OMAP4 -
Cortex-A9, Dual Core
External L2 cache ( PL310)
New Interrupt controller(GIC)
New Memory subsystem (EMIF)
3 Additional co-procesors
New DSS
IVHA2

Quoted 1 line

Except ARMv7 architecture, there is not much common between these two SOCs

Thanks for the clarification.

Also, as Russell pointed out (which I didn't know), you can't boot an SMP kernel on a UP machine. So I now understand the significance of your original SMP build point.

#9 lore raw
In reply to Simon Horman
Quoted 3 lines · Simon Horman

Are there any other reasons to keep omap_4430sdp_defconfig? SMP seems to be a weak argument to me, after all if I peek in to arch/x86/configs/ I don't see SMP and non-SMP variants.

SMP is a strong argument on ARM - a SMP kernel can't be booted on UP machines, whereas on x86, a SMP kernel can be booted on UP machines.

#10 lore raw
Quoted 7 lines · Russell King - ARM Linux
Quoted 3 lines · Simon Horman

Are there any other reasons to keep omap_4430sdp_defconfig? SMP seems to be a weak argument to me, after all if I peek in to arch/x86/configs/ I don't see SMP and non-SMP variants.

SMP is a strong argument on ARM - a SMP kernel can't be booted on UP machines, whereas on x86, a SMP kernel can be booted on UP machines.

Late ack from me to stop building other omap defconfigs than omap3_defconfig and omap_4430sdp_defconfig.

Cheers,

Tony

#11 lore raw

PATCH 1/1 ARM: ns9xxx: fix FTBFS for zImage

In reply to Vincent Sanders

the different putc variants used an initialized local static variable which is broken since

5de813b (ARM: Eliminate decompressor -Dstatic= PIC hack)

This needs to be initialized at runtime and so needs to be global. While at it give it a better name.

Signed-off-by Uwe Kleine-K?nig <[email protected]>

---
KernelVersion: v2.6.35-rc6

 arch/arm/mach-ns9xxx/include/mach/uncompress.h |   30 ++++++++++++------------
 1 files changed, 15 insertions(+), 15 deletions(-)
arch/arm/mach-ns9xxx/include/mach/uncompress.h +15 -15
index 1b12d32..770a68c 100644
@@ -20,50 +20,49 @@ static void putc_dummy(char c, void __iomem *base)
 	/* nothing */
 }
 
+static int timeout;
+
 static void putc_ns9360(char c, void __iomem *base)
 {
-	static int t = 0x10000;
 	do {
-		if (t)
-			--t;
+		if (timeout)
+			--timeout;
 
 		if (__raw_readl(base + 8) & (1 << 3)) {
 			__raw_writeb(c, base + 16);
-			t = 0x10000;
+			timeout = 0x10000;
 			break;
 		}
-	} while (t);
+	} while (timeout);
 }
 
 static void putc_a9m9750dev(char c, void __iomem *base)
 {
-	static int t = 0x10000;
 	do {
-		if (t)
-			--t;
+		if (timeout)
+			--timeout;
 
 		if (__raw_readb(base + 5) & (1 << 5)) {
 			__raw_writeb(c, base);
-			t = 0x10000;
+			timeout = 0x10000;
 			break;
 		}
-	} while (t);
+	} while (timeout);
 
 }
 
 static void putc_ns921x(char c, void __iomem *base)
 {
-	static int t = 0x10000;
 	do {
-		if (t)
-			--t;
+		if (timeout)
+			--timeout;
 
 		if (!(__raw_readl(base) & (1 << 11))) {
 			__raw_writeb(c, base + 0x0028);
-			t = 0x10000;
+			timeout = 0x10000;
 			break;
 		}
-	} while (t);
+	} while (timeout);
 }
 
 #define MSCS __REG(0xA0900184)

@@ -89,6 +88,7 @@ static void putc_ns921x(char c, void __iomem *base)
 
 static void autodetect(void (**putc)(char, void __iomem *), void __iomem **base)
 {
+	timeout = 0x10000;
 	if (((__raw_readl(MSCS) >> 16) & 0xfe) == 0x00) {
 		/* ns9360 or ns9750 */
 		if (NS9360_UART_ENABLED(NS9360_UARTA)) {
Signature · 1 line
1.7.1
#12 lore raw

PATCH 1/1 ARM: ns9xxx: fix FTBFS for zImage

Subject: [PATCH] ARM: ns9xxx: fix FTBFS for zImage

btw, I don't think you mean FTBFS here.

"fix fails to build from source for zImage"

is wrong.

"fix build failure for zImage"

is better.

#13 lore raw

PATCH 1/1 ARM: ns9xxx: fix FTBFS for zImage

Quoted 11 lines · Russell King - ARM Linux

Subject: [PATCH] ARM: ns9xxx: fix FTBFS for zImage

btw, I don't think you mean FTBFS here.

"fix fails to build from source for zImage"

is wrong.

"fix build failure for zImage"

is better.

I didn't thought much about it, but if F means "failure" it should be OK :-).

Best regards
Uwe

Signature · 2 lines
Pengutronix e.K.                           | Uwe Kleine-K?nig            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
#14 lore raw
In reply to Vincent Sanders

Hello Vincent,

Quoted 2 lines · Vincent Sanders

The ns9xxx configuration has been failing for even longer and teh machines supported are maintainerd by Uwe Kleine-Konig

See patch that I just sent to this thread. Russell: is it already to late for .36? Should I send my patch to your patch system? There is also still a Kconfig-Warning in some imx-builds.

Vincent: one thing I noticed a while back is that kautobuild doesn't report the section mismatch warnings. See for example

	http://armlinux.simtec.co.uk/kautobuild/2.6.35-rc6/cam60_defconfig/zimage.log

It has:

...
LD vmlinux.o
MODPOST vmlinux.o
WARNING: modpost: Found 1 section mismatch(es). To see full details build your kernel with:
'make CONFIG_DEBUG_SECTION_MISMATCH=y'

Still

	http://armlinux.simtec.co.uk/kautobuild/2.6.35-rc6/cam60_defconfig.html

claims to have found no warnings.

Can you please consider adding CONFIG_DEBUG_SECTION_MISMATCH=y to your build and extend the warning regexp?

Best regards
Uwe

Signature · 2 lines
Pengutronix e.K.                           | Uwe Kleine-K?nig            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
#15 lore raw
Quoted 7 lines · Uwe Kleine-K?nig

Hello Vincent,

Quoted 2 lines · Vincent Sanders

The ns9xxx configuration has been failing for even longer and teh machines supported are maintainerd by Uwe Kleine-Konig

See patch that I just sent to this thread. Russell: is it already to late for .36?

No, I still have patches queued to go into the -rc series.

Quoted 1 line

Should I send my patch to your patch system?

Yes please.