sparc: don't #include asm/system.h in asm/jump_label.h
Subsystems (2)
PATCH 1/3 sparc: don't #include asm/system.h in asm/jump_label.h
It seems that #include <asm/system.h> makes a circular dependency between kernel.h and bitmap.h which breaks allmodconfig build. Removing the line makes no change because jump_label.h doesn't need it actually AFAICS. Compile tested on sparc32 allmodconfig.
---
Failed build log is as following:
CC arch/sparc/kernel/asm-offsets.s
In file included from include/linux/cpumask.h:11:0,
from arch/sparc/include/asm/smp_32.h:15,
from arch/sparc/include/asm/smp.h:6,
from arch/sparc/include/asm/system_32.h:12,
from arch/sparc/include/asm/system.h:6,
from arch/sparc/include/asm/jump_label.h:7,
from include/linux/jump_label.h:5,
from include/linux/dynamic_debug.h:4,
from include/linux/kernel.h:20,
from include/linux/sched.h:54,
from arch/sparc/kernel/asm-offsets.c:13:
include/linux/bitmap.h: In function 'bitmap_zero':
include/linux/bitmap.h:159:3: error: implicit declaration of function 'DIV_ROUND_UP'
In file included from arch/sparc/include/asm/smp_32.h:15:0,
from arch/sparc/include/asm/smp.h:6,
from arch/sparc/include/asm/system_32.h:12,
from arch/sparc/include/asm/system.h:6,
from arch/sparc/include/asm/jump_label.h:7,
from include/linux/jump_label.h:5,
from include/linux/dynamic_debug.h:4,
from include/linux/kernel.h:20,
from include/linux/sched.h:54,
from arch/sparc/kernel/asm-offsets.c:13:include/linux/cpumask.h: At top level:include/linux/cpumask.h:13:26: error: variably modified 'bits' at file scope include/linux/cpumask.h:659:14: error: variably modified 'cpu_all_bits' at file scope include/linux/cpumask.h:705:2: error: variably modified 'cpu_bit_bitmap' at file scope make[3]: *** [arch/sparc/kernel/asm-offsets.s] Error 1
make[2]: *** [prepare0] Error 2
make[1]: *** [sub-make] Error 2
make: *** [all] Error 2
arch/sparc/include/asm/jump_label.h | 1 - 1 files changed, 0 insertions(+), 1 deletions(-)
arch/sparc/include/asm/jump_label.h +0 -1
@@ -4,7 +4,6 @@#ifdef __KERNEL__#include <linux/types.h> -#include <asm/system.h> #define JUMP_LABEL_NOP_SIZE 4
Signature · 1 line
1.7.0.4
PATCH 2/3 sparc32: remove CONFIG_HAVE_PERF_EVENTS option
Remove HAVE_PERF_EVENTS and PERF_USE_VMALLOC under config SPARC because they're under SPARC64 too. Supporting perf_event needs atomic64 operations but AFAIK sparc32 doesn't provide them, CMIIW. ;-) Also removes redundant HAVE_IRQ_WORK line.
--- arch/sparc/Kconfig | 3 --- 1 files changed, 0 insertions(+), 3 deletions(-)
arch/sparc/Kconfig +0 -3
@@ -27,8 +27,6 @@ config SPARC select RTC_CLASS select RTC_DRV_M48T59 select HAVE_IRQ_WORK - select HAVE_PERF_EVENTS - select PERF_USE_VMALLOC select HAVE_DMA_ATTRS select HAVE_DMA_API_DEBUG select HAVE_ARCH_JUMP_LABEL
@@ -55,7 +53,6 @@ config SPARC64 select RTC_DRV_BQ4802 select RTC_DRV_SUN4V select RTC_DRV_STARFIRE - select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select PERF_USE_VMALLOC
Signature · 1 line
1.7.0.4
David Miller
maintainerRe: [PATCH 2/3] sparc32: remove CONFIG_HAVE_PERF_EVENTS option
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 00:52:37 +0900
Quoted 7 lines
Remove
HAVE_PERF_EVENTSandPERF_USE_VMALLOCunder config SPARC because they're under SPARC64 too. Supportingperf_eventneeds atomic64 operations but AFAIK sparc32 doesn't provide them, CMIIW. ;-) Also removes redundantHAVE_IRQ_WORKline.Signed-off-by Namhyung Kim <[email protected]>
The generic non-hw perf event stuff builds fine on sparc32, please don't remove this unless you have a legitimate build failure to report and fix.
Even if sparc32 doesn't provide atomic64_t support, it very easily could do so using include/asm-generic/atomic64.h which implements them using spinlocks.
Re: [PATCH 2/3] sparc32: remove CONFIG_HAVE_PERF_EVENTS option
2010-10-25 (ì›”), 08:59 -0700, David Miller:
Quoted 11 lines
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 00:52:37 +0900Quoted 7 lines
Remove
HAVE_PERF_EVENTSandPERF_USE_VMALLOCunder config SPARC because they're under SPARC64 too. Supportingperf_eventneeds atomic64 operations but AFAIK sparc32 doesn't provide them, CMIIW. ;-) Also removes redundantHAVE_IRQ_WORKline.Signed-off-by Namhyung Kim <[email protected]>
Quoted 7 lines
The generic non-hw perf event stuff builds fine on sparc32, please don't remove this unless you have a legitimate build failure to report and fix.
Even if sparc32 doesn't provide
atomic64_tsupport, it very easily could do so usinginclude/asm-generic/atomic64.hwhich implements them using spinlocks.
Now I see following build failure without this patch. Maybe some work is needed to deal with atomic64_t. I'll try to prepare for it soon. Thanks.
CC arch/sparc/kernel/setup_32.o
In file included from arch/sparc/include/asm/local64.h:1:0,
from include/linux/perf_event.h:465,
from include/linux/ftrace_event.h:8,
from include/trace/syscall.h:6,
from include/linux/syscalls.h:76,
from arch/sparc/kernel/setup_32.c:23:
include/asm-generic/local64.h:62:2: error: expected
specifier-qualifier-list before 'atomic64_t'
In file included from include/linux/ftrace_event.h:8:0,
from include/trace/syscall.h:6,
from include/linux/syscalls.h:76,
from arch/sparc/kernel/setup_32.c:23:include/linux/perf_event.h:722:2: error: expected
specifier-qualifier-list before 'atomic64_t'include/linux/perf_event.h: In function 'perf_fetch_caller_regs':include/linux/perf_event.h:1004:2: error: 'struct pt_regs' has no member
named 'tstate'include/linux/perf_event.h:1004:2: error: 'struct pt_regs' has no member
named 'tpc'include/linux/perf_event.h:1004:2: error: 'struct pt_regs' has no member
named 'tnpc'include/linux/perf_event.h:1004:2: error: 'struct pt_regs' has no member
named 'tpc'
make[3]: *** [arch/sparc/kernel/setup_32.o] Error 1
make[2]: *** [arch/sparc/kernel] Error 2
make[1]: *** [sub-make] Error 2
make: *** [all] Error 2
Signature · 2 lines
Regards, Namhyung Kim
Re: [PATCH 2/3] sparc32: remove CONFIG_HAVE_PERF_EVENTS option
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 01:31:19 +0900
Quoted 16 lines
2010-10-25 (월), 08:59 -0700, David Miller:
Quoted 15 lines
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 00:52:37 +0900Quoted 7 lines
Remove
HAVE_PERF_EVENTSandPERF_USE_VMALLOCunder config SPARC because they're under SPARC64 too. Supportingperf_eventneeds atomic64 operations but AFAIK sparc32 doesn't provide them, CMIIW. ;-) Also removes redundantHAVE_IRQ_WORKline.Signed-off-by Namhyung Kim <[email protected]>The generic non-hw perf event stuff builds fine on sparc32, please don't remove this unless you have a legitimate build failure to report and fix.
Quoted 6 lines
Quoted 3 lines
Even if sparc32 doesn't provide
atomic64_tsupport, it very easily could do so usinginclude/asm-generic/atomic64.hwhich implements them using spinlocks.Now I see following build failure without this patch. Maybe some work is needed to deal with
atomic64_t. I'll try to prepare for it soon. Thanks.
Nevermind, I see the issue. Ok, your patch is fine:
PATCH 3/3 sparc32: fix build failure on CONFIG_SPARC_LEON
In reply to Namhyung Kim CC arch/sparc/kernel/irq_32.oarch/sparc/kernel/irq_32.c: In function 'request_fast_irq':arch/sparc/kernel/irq_32.c:370:25: error: conflicting types for 'trapbase_cpu1' arch/sparc/include/asm/leon.h:366:22: note: previous declaration of 'trapbase_cpu1' was here arch/sparc/kernel/irq_32.c:370:40: error: conflicting types for 'trapbase_cpu2' arch/sparc/include/asm/leon.h:367:22: note: previous declaration of 'trapbase_cpu2' was here arch/sparc/kernel/irq_32.c:370:55: error: conflicting types for 'trapbase_cpu3' arch/sparc/include/asm/leon.h:368:22: note: previous declaration of 'trapbase_cpu3' was here make[3]: *** [arch/sparc/kernel/irq_32.o] Error 1
make[2]: *** [arch/sparc/kernel] Error 2
make[1]: *** [sub-make] Error 2
make: *** [all] Error 2
--- arch/sparc/kernel/irq_32.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-)
arch/sparc/kernel/irq_32.c +2 -2
@@ -365,7 +365,7 @@ static int request_fast_irq(unsigned int irq,unsigned long flags;unsigned int cpu_irq;int ret; -#ifdef CONFIG_SMP +#if definedCONFIG_SMP && !definedCONFIG_SPARC_LEON struct tt_entry *trap_table;extern struct tt_entry trapbase_cpu1, trapbase_cpu2, trapbase_cpu3;#endif
@@ -425,7 +425,7 @@ static int request_fast_irq(unsigned int irq, table[SP_TRAP_IRQ1 +(cpu_irq-1 )].inst_four =SPARC_NOP ; INSTANTIATE(sparc_ttable) -#ifdef CONFIG_SMP +#if definedCONFIG_SMP && !definedCONFIG_SPARC_LEON trap_table = &trapbase_cpu1;INSTANTIATE (trap_table) trap_table = &trapbase_cpu2;INSTANTIATE (trap_table) trap_table = &trapbase_cpu3;INSTANTIATE (trap_table)
Signature · 1 line
1.7.0.4
David Miller
maintainerRe: [PATCH 3/3] sparc32: fix build failure on CONFIG_SPARC_LEON
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 00:52:38 +0900
Quoted 3 lines
CC
arch/sparc/kernel/irq_32.oarch/sparc/kernel/irq_32.c: In function 'request_fast_irq':arch/sparc/kernel/irq_32.c:370:25: error: conflicting types for 'trapbase_cpu1'
I doubt SMP LEON chips even exist, so better to just anti-dep this in arch/sparc/Kconfig
Re: [PATCH 3/3] sparc32: fix build failure on CONFIG_SPARC_LEON
2010-10-25 (ì›”), 09:00 -0700, David Miller:
Quoted 9 lines
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 00:52:38 +0900Quoted 3 lines
CC
arch/sparc/kernel/irq_32.oarch/sparc/kernel/irq_32.c: In function 'request_fast_irq':arch/sparc/kernel/irq_32.c:370:25: error: conflicting types for 'trapbase_cpu1'I doubt SMP LEON chips even exist, so better to just anti-dep this in
arch/sparc/Kconfig
If I understood correctly, you meant this?
arch/sparc/Kconfig +1 -1
@@ -463,7 +463,7 @@ config SERIAL_CONSOLE config SPARC_LEON bool "Sparc Leon processor family" - depends on SPARC32 + depends on SPARC32 && !SMP ---help--- If you say Y here if you are running on a SPARC-LEON
processor.
The LEON processor is a synthesizable VHDL model of the
Signature · 2 lines
Regards, Namhyung Kim
David Miller
maintainerRe: [PATCH 3/3] sparc32: fix build failure on CONFIG_SPARC_LEON
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 01:07:29 +0900
Quoted 4 lines
2010-10-25 (월), 09:00 -0700, David Miller:
Quoted 3 lines
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 00:52:38 +0900
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
CC
arch/sparc/kernel/irq_32.oarch/sparc/kernel/irq_32.c: In function 'request_fast_irq':arch/sparc/kernel/irq_32.c:370:25: error: conflicting types for 'trapbase_cpu1'I doubt SMP LEON chips even exist, so better to just anti-dep this in
arch/sparc/KconfigIf I understood correctly, you meant this?
Yes, something like that.
Re: [PATCH 3/3] sparc32: fix build failure on CONFIG_SPARC_LEON
In reply to David MillerQuoted 9 lines · David Miller
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 00:52:38 +0900Quoted 3 lines
CC
arch/sparc/kernel/irq_32.oarch/sparc/kernel/irq_32.c: In function 'request_fast_irq':arch/sparc/kernel/irq_32.c:370:25: error: conflicting types for 'trapbase_cpu1'I doubt SMP LEON chips even exist, so better to just anti-dep this in
arch/sparc/Kconfig
I have a nice eval board at my desk with a dual-core LEON4 processor. I only lack time to play with it...
Sam
David Miller
maintainerRe: [PATCH 3/3] sparc32: fix build failure on CONFIG_SPARC_LEON
From: Sam Ravnborg <[email protected]>
Date: Mon, 25 Oct 2010 18:11:22 +0200
Quoted 8 lines · David Miller
Quoted 7 lines
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 00:52:38 +0900Quoted 3 lines
CC
arch/sparc/kernel/irq_32.oarch/sparc/kernel/irq_32.c: In function 'request_fast_irq':arch/sparc/kernel/irq_32.c:370:25: error: conflicting types for 'trapbase_cpu1'
Quoted 5 lines
Quoted 2 lines
I doubt SMP LEON chips even exist, so better to just anti-dep this in
arch/sparc/KconfigI have a nice eval board at my desk with a dual-core LEON4 processor. I only lack time to play with it...
Ok. But whenever someone gets a time to play with it and verify it works, we can sort out this build failure by modifying the code as needed.
But for now I prefer the Kconfig SMP dependency fix.
Re: [PATCH 3/3] sparc32: fix build failure on CONFIG_SPARC_LEON
Quoted 14 lines · David Miller
From: Sam Ravnborg <[email protected]>
Date: Mon, 25 Oct 2010 18:11:22 +0200Quoted 11 lines · David Miller
Quoted 9 lines
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 00:52:38 +0900Quoted 3 lines
CC
arch/sparc/kernel/irq_32.oarch/sparc/kernel/irq_32.c: In function 'request_fast_irq':arch/sparc/kernel/irq_32.c:370:25: error: conflicting types for 'trapbase_cpu1'I doubt SMP LEON chips even exist, so better to just anti-dep this in
arch/sparc/Kconfig
Quoted 8 lines
Quoted 2 lines
I have a nice eval board at my desk with a dual-core LEON4 processor. I only lack time to play with it...
Ok. But whenever someone gets a time to play with it and verify it works, we can sort out this build failure by modifying the code as needed.
But for now I prefer the
KconfigSMP dependency fix.
Yup - I agree.
My only intention was to tell that such a processor exists in reality and not only on paper.
Sam
Re: [PATCH 3/3] sparc32: fix build failure on CONFIG_SPARC_LEON
Quoted 33 lines · Sam Ravnborg
Quoted 30 lines · David Miller
From: Sam Ravnborg <[email protected]>
Date: Mon, 25 Oct 2010 18:11:22 +0200Quoted 13 lines · David Miller
Quoted 12 lines
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 00:52:38 +0900Quoted 5 lines
CC
arch/sparc/kernel/irq_32.oarch/sparc/kernel/irq_32.c: In function 'request_fast_irq':arch/sparc/kernel/irq_32.c:370:25: error: conflicting types for 'trapbase_cpu1'Quoted 8 lines
Quoted 4 lines
I doubt SMP LEON chips even exist, so better to just anti-dep this in
arch/sparc/KconfigI have a nice eval board at my desk with a dual-core LEON4 processor. I only lack time to play with it...
Ok. But whenever someone gets a time to play with it and verify it works, we can sort out this build failure by modifying the code as needed.
Quoted 8 lines
Quoted 3 lines
But for now I prefer the
KconfigSMP dependency fix.Yup - I agree.
My only intention was to tell that such a processor exists in reality and not only on paper.
I stumbled into the same problem last weeks, enabling "Normal floppy disk support" (BLK_DEV_FD) on the LEON makes this error appear. There is no hardware for "Normal floppy disk support" on a LEON.
Regards,
Daniel Hellstrom
From: Namhyung Kim <[email protected]>
Date: Tue, 26 Oct 2010 00:52:36 +0900
Quoted 6 lines
It seems that #include <
asm/system.h> makes a circular dependency betweenkernel.handbitmap.hwhich breaks allmodconfig build. Removing the line makes no change becausejump_label.hdoesn't need it actually AFAICS. Compile tested on sparc32 allmodconfig.Signed-off-by Namhyung Kim <[email protected]>