[v2][PATCH 1/3] kgdb,ppc: do not set kgdb_single_step on ppc

[v2][PATCH 1/3] kgdb,ppc: do not set kgdb_single_step on ppc

[Tiejun Chen]

The kgdb_single_step flag has the possibility to indefinitely
hang the system on an SMP system.

The x86 arch have the same problem, and that problem was fixed by
commit 8097551d9ab9b9e3630(kgdb,x86: do not set kgdb_single_step
on x86). This patch does the same behaviors as x86's patch.

Signed-off-by: Dongdong Deng <dongdong.deng@windriver.com>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
---
v2: nothing changed.

 arch/powerpc/kernel/kgdb.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

diff --git a/arch/powerpc/kernel/kgdb.c b/arch/powerpc/kernel/kgdb.c
index 782bd0a..bbabc5a 100644
--- a/arch/powerpc/kernel/kgdb.c
+++ b/arch/powerpc/kernel/kgdb.c
@@ -410,7 +410,6 @@ int kgdb_arch_handle_exception(int vector, int signo, int err_code,
 #else
 			linux_regs->msr |= MSR_SE;
 #endif
-			kgdb_single_step = 1;
 			atomic_set(&kgdb_cpu_doing_single_step,
 				   raw_smp_processor_id());
 		}
-- 
1.5.6

[v2][PATCH 2/3] powerpc: Bail out of KGDB when we've been triggered

[Tiejun Chen]

We need to skip a breakpoint exception when it occurs after
a breakpoint has already been removed.

Signed-off-by: Tiejun Chen <tiejun.chen@windriver.com>
---
v2: simply kgdb_skipexception() return path. 

 arch/powerpc/kernel/kgdb.c |   15 +++++++++++++++
 1 files changed, 15 insertions(+), 0 deletions(-)

diff --git a/arch/powerpc/kernel/kgdb.c b/arch/powerpc/kernel/kgdb.c
index bbabc5a..05adb69 100644
--- a/arch/powerpc/kernel/kgdb.c
+++ b/arch/powerpc/kernel/kgdb.c
@@ -101,6 +101,21 @@ static int computeSignal(unsigned int tt)
 	return SIGHUP;		/* default for things we don't know about */
 }
 
+/**
+ *
+ *	kgdb_skipexception - Bail out of KGDB when we've been triggered.
+ *	@exception: Exception vector number
+ *	@regs: Current &struct pt_regs.
+ *
+ *	On some architectures we need to skip a breakpoint exception when
+ *	it occurs after a breakpoint has been removed.
+ *
+ */
+int kgdb_skipexception(int exception, struct pt_regs *regs)
+{
+	return kgdb_isremovedbreak(regs->nip);
+}
+
 static int kgdb_call_nmi_hook(struct pt_regs *regs)
 {
 	kgdb_nmicallback(raw_smp_processor_id(), regs);
-- 
1.5.6

[v2][PATCH 3/3] powerpc/kgdb: restore current_thread_info properly

[Tiejun Chen]

For powerpc BooKE and e200, singlestep is handled on the critical/dbg
exception stack. This causes current_thread_info() to fail for kgdb
internal, so previously We work around this issue by copying
the thread_info from the kernel stack before calling kgdb_handle_exception,
and copying it back afterwards.

But actually we don't do this properly. We should backup current_thread_info
then restore that when exit.

Signed-off-by: Tiejun Chen <tiejun.chen@windriver.com>
---
v2: fix a typo in patch head description.

 arch/powerpc/kernel/kgdb.c |   11 +++++++++--
 1 files changed, 9 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/kernel/kgdb.c b/arch/powerpc/kernel/kgdb.c
index 05adb69..c470a40 100644
--- a/arch/powerpc/kernel/kgdb.c
+++ b/arch/powerpc/kernel/kgdb.c
@@ -25,6 +25,7 @@
 #include <asm/processor.h>
 #include <asm/machdep.h>
 #include <asm/debug.h>
+#include <linux/slab.h>
 
 /*
  * This table contains the mapping between PowerPC hardware trap types, and
@@ -153,6 +154,8 @@ static int kgdb_handle_breakpoint(struct pt_regs *regs)
 static int kgdb_singlestep(struct pt_regs *regs)
 {
 	struct thread_info *thread_info, *exception_thread_info;
+	struct thread_info *backup_current_thread_info = \
+		(struct thread_info *)kmalloc(sizeof(struct thread_info), GFP_KERNEL);
 
 	if (user_mode(regs))
 		return 0;
@@ -170,13 +173,17 @@ static int kgdb_singlestep(struct pt_regs *regs)
 	thread_info = (struct thread_info *)(regs->gpr[1] & ~(THREAD_SIZE-1));
 	exception_thread_info = current_thread_info();
 
-	if (thread_info != exception_thread_info)
+	if (thread_info != exception_thread_info) {
+		/* Save the original current_thread_info. */
+		memcpy(backup_current_thread_info, exception_thread_info, sizeof *thread_info);
 		memcpy(exception_thread_info, thread_info, sizeof *thread_info);
+	}
 
 	kgdb_handle_exception(0, SIGTRAP, 0, regs);
 
 	if (thread_info != exception_thread_info)
-		memcpy(thread_info, exception_thread_info, sizeof *thread_info);
+		/* Restore current_thread_info lastly. */
+		memcpy(exception_thread_info, backup_current_thread_info, sizeof *thread_info);
 
 	return 1;
 }
-- 
1.5.6

Re: [v2][PATCH 3/3] powerpc/kgdb: restore current_thread_info properly

[Nicholas A. Bellinger]

On Thu, 2012-08-23 at 10:10 +0800, Tiejun Chen wrote:
> For powerpc BooKE and e200, singlestep is handled on the critical/dbg
> exception stack. This causes current_thread_info() to fail for kgdb
> internal, so previously We work around this issue by copying
> the thread_info from the kernel stack before calling kgdb_handle_exception,
> and copying it back afterwards.
> 
> But actually we don't do this properly. We should backup current_thread_info
> then restore that when exit.
> 
> Signed-off-by: Tiejun Chen <tiejun.chen@windriver.com>
> ---
> v2: fix a typo in patch head description.
> 
>  arch/powerpc/kernel/kgdb.c |   11 +++++++++--
>  1 files changed, 9 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/powerpc/kernel/kgdb.c b/arch/powerpc/kernel/kgdb.c
> index 05adb69..c470a40 100644
> --- a/arch/powerpc/kernel/kgdb.c
> +++ b/arch/powerpc/kernel/kgdb.c
> @@ -25,6 +25,7 @@
>  #include <asm/processor.h>
>  #include <asm/machdep.h>
>  #include <asm/debug.h>
> +#include <linux/slab.h>
>  
>  /*
>   * This table contains the mapping between PowerPC hardware trap types, and
> @@ -153,6 +154,8 @@ static int kgdb_handle_breakpoint(struct pt_regs *regs)
>  static int kgdb_singlestep(struct pt_regs *regs)
>  {
>  	struct thread_info *thread_info, *exception_thread_info;
> +	struct thread_info *backup_current_thread_info = \
> +		(struct thread_info *)kmalloc(sizeof(struct thread_info), GFP_KERNEL);
>  

Looks like a rouge '\' in the above assignment..

Re: [v2][PATCH 3/3] powerpc/kgdb: restore current_thread_info properly

[tiejun.chen]

On 08/23/2012 11:14 AM, Nicholas A. Bellinger wrote:
> On Thu, 2012-08-23 at 10:10 +0800, Tiejun Chen wrote:
>> For powerpc BooKE and e200, singlestep is handled on the critical/dbg
>> exception stack. This causes current_thread_info() to fail for kgdb
>> internal, so previously We work around this issue by copying
>> the thread_info from the kernel stack before calling kgdb_handle_exception,
>> and copying it back afterwards.
>>
>> But actually we don't do this properly. We should backup current_thread_info
>> then restore that when exit.
>>
>> Signed-off-by: Tiejun Chen <tiejun.chen@windriver.com>
>> ---
>> v2: fix a typo in patch head description.
>>
>>  arch/powerpc/kernel/kgdb.c |   11 +++++++++--
>>  1 files changed, 9 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/powerpc/kernel/kgdb.c b/arch/powerpc/kernel/kgdb.c
>> index 05adb69..c470a40 100644
>> --- a/arch/powerpc/kernel/kgdb.c
>> +++ b/arch/powerpc/kernel/kgdb.c
>> @@ -25,6 +25,7 @@
>>  #include <asm/processor.h>
>>  #include <asm/machdep.h>
>>  #include <asm/debug.h>
>> +#include <linux/slab.h>
>>  
>>  /*
>>   * This table contains the mapping between PowerPC hardware trap types, and
>> @@ -153,6 +154,8 @@ static int kgdb_handle_breakpoint(struct pt_regs *regs)
>>  static int kgdb_singlestep(struct pt_regs *regs)
>>  {
>>  	struct thread_info *thread_info, *exception_thread_info;
>> +	struct thread_info *backup_current_thread_info = \
>> +		(struct thread_info *)kmalloc(sizeof(struct thread_info), GFP_KERNEL);
>>  
> 
> Looks like a rouge '\' in the above assignment..

Remove that :)

Thanks
Tiejun