[PATCH] xen/arm: traps: Print a message in debug build when a guest dabt is not handled

[PATCH] xen/arm: traps: Print a message in debug build when a guest dabt is not handled

[Julien Grall]

This is useful for debugging low level kernel before the guest as setup
the vector table.

Note that the value of the IPA is only here for reference and may not
always be valid if the error came from a stage 1 table translation walk.

Signed-off-by: Julien Grall <julien.grall@citrix.com>
---

Rather than having a check only for UART, a more generic message would
be more useful in debug build.
---
 xen/arch/arm/traps.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/xen/arch/arm/traps.c b/xen/arch/arm/traps.c
index 47d6cef..1dcc31f 100644
--- a/xen/arch/arm/traps.c
+++ b/xen/arch/arm/traps.c
@@ -2355,7 +2355,6 @@ static void do_trap_data_abort_guest(struct cpu_user_regs *regs,
 #else
     info.gva = READ_SYSREG64(FAR_EL2);
 #endif
-
     if ( dabt.s1ptw )
         info.gpa = READ_SYSREG(HPFAR_EL2);
     else
@@ -2413,6 +2412,8 @@ static void do_trap_data_abort_guest(struct cpu_user_regs *regs,
     }
 
 bad_data_abort:
+    gdprintk(XENLOG_DEBUG, "HSR=0x%x pc=%#"PRIregister" gva=%#"PRIvaddr
+             " gpa=%#"PRIpaddr"\n", hsr.bits, regs->pc, info.gva, info.gpa);
     inject_dabt_exception(regs, info.gva, hsr.len);
 }
 
-- 
2.1.4

Re: [PATCH] xen/arm: traps: Print a message in debug build when a guest dabt is not handled

[Ian Campbell]

On Thu, 2015-05-28 at 10:10 +0100, Julien Grall wrote:
> This is useful for debugging low level kernel before the guest as setup
> the vector table.
> 
> Note that the value of the IPA is only here for reference and may not
> always be valid if the error came from a stage 1 table translation walk.
> 
> Signed-off-by: Julien Grall <julien.grall@citrix.com>

Acked + applied, but dropping the spurious whitespace change, since it
didn't look deliberate.

Re: [PATCH] xen/arm: traps: Print a message in debug build when a guest dabt is not handled

[Julien Grall]

On 03/06/2015 11:34, Ian Campbell wrote:
> On Thu, 2015-05-28 at 10:10 +0100, Julien Grall wrote:
>> This is useful for debugging low level kernel before the guest as setup
>> the vector table.
>>
>> Note that the value of the IPA is only here for reference and may not
>> always be valid if the error came from a stage 1 table translation walk.
>>
>> Signed-off-by: Julien Grall <julien.grall@citrix.com>
>
> Acked + applied, but dropping the spurious whitespace change, since it
> didn't look deliberate.

Right, I did some debugging to check the validity of this patch but 
forgot to remove all the lines.

Thank you,

-- 
Julien Grall