2017-12-11 16:06:01

by Dongjiu Geng

Subject: [RESEND PATCH V2] arm64: fault: avoid send SIGBUS two times

do_sea() calls arm64_notify_die() which will always signal
user-space. It also returns whether APEI claimed the external
abort as a RAS notification. If it returns failure do_mem_abort()
will signal user-space too.

do_mem_abort() wants to know if we handled the error, we always
call arm64_notify_die() so can always return success.

Signed-off-by: Dongjiu Geng <[email protected]>
---
1. Address James's comments to update the commit messages
2. Address James's comments to not change the si_code for SIGBUS
---
arch/arm64/mm/fault.c | 7 +++----
1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
index b64958b..38b9f3e 100644
--- a/arch/arm64/mm/fault.c
+++ b/arch/arm64/mm/fault.c
@@ -610,7 +610,6 @@ static int do_sea(unsigned long addr, unsigned int esr, struct pt_regs *regs)
{
struct siginfo info;
const struct fault_info *inf;
- int ret = 0;

inf = esr_to_fault_info(esr);
pr_err("Synchronous External Abort: %s (0x%08x) at 0x%016lx\n",
@@ -625,7 +624,7 @@ static int do_sea(unsigned long addr, unsigned int esr, struct pt_regs *regs)
if (interrupts_enabled(regs))
nmi_enter();

- ret = ghes_notify_sea();
+ ghes_notify_sea();

if (interrupts_enabled(regs))
nmi_exit();
@@ -640,7 +639,7 @@ static int do_sea(unsigned long addr, unsigned int esr, struct pt_regs *regs)
info.si_addr = (void __user *)addr;
arm64_notify_die("", regs, &info, esr);

- return ret;
+ return 0;
}

static const struct fault_info fault_info[] = {
--
2.10.1

2017-12-12 03:31:43

by Xie XiuQi

Subject: Re: [RESEND PATCH V2] arm64: fault: avoid send SIGBUS two times

Hi Dongjiu,

On 2017/12/12 0:05, Dongjiu Geng wrote:
> do_sea() calls arm64_notify_die() which will always signal
> user-space. It also returns whether APEI claimed the external
> abort as a RAS notification. If it returns failure do_mem_abort()
> will signal user-space too.
>
> do_mem_abort() wants to know if we handled the error, we always
> call arm64_notify_die() so can always return success.
>
> Signed-off-by: Dongjiu Geng <[email protected]>
> ---
> 1. Address James's comments to update the commit messages
> 2. Address James's comments to not change the si_code for SIGBUS
> ---
> arch/arm64/mm/fault.c | 7 +++----
> 1 file changed, 3 insertions(+), 4 deletions(-)
>
> diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
> index b64958b..38b9f3e 100644
> --- a/arch/arm64/mm/fault.c
> +++ b/arch/arm64/mm/fault.c
> @@ -610,7 +610,6 @@ static int do_sea(unsigned long addr, unsigned int esr, struct pt_regs *regs)
> {
> struct siginfo info;
> const struct fault_info *inf;
> - int ret = 0;
>
> inf = esr_to_fault_info(esr);
> pr_err("Synchronous External Abort: %s (0x%08x) at 0x%016lx\n",
> @@ -625,7 +624,7 @@ static int do_sea(unsigned long addr, unsigned int esr, struct pt_regs *regs)
> if (interrupts_enabled(regs))
> nmi_enter();
>
> - ret = ghes_notify_sea();
> + ghes_notify_sea();
>
> if (interrupts_enabled(regs))
> nmi_exit();
> @@ -640,7 +639,7 @@ static int do_sea(unsigned long addr, unsigned int esr, struct pt_regs *regs)
> info.si_addr = (void __user *)addr;
> arm64_notify_die("", regs, &info, esr);
>
> - return ret;
> + return 0;

It looks good to me. do_sea() has done all necessary action for SEA, so it should always return 0,
no matter ghes_notify_sea() return true or false.

Reviewed-by: Xie XiuQi <[email protected]>

> }
>
> static const struct fault_info fault_info[] = {
> --
> 2.10.1
>

--
Thanks,
Xie XiuQi

2017-12-12 04:15:33

by Dongjiu Geng

Subject: Re: [RESEND PATCH V2] arm64: fault: avoid send SIGBUS two times

On 2017/12/12 11:31, Xie XiuQi wrote:
>> + return 0;
> It looks good to me. do_sea() has done all necessary action for SEA, so it should always return 0,
> no matter ghes_notify_sea() return true or false.
yes, it is.

>
> Reviewed-by: Xie XiuQi <[email protected]>

Thanks XiuQi's review and comments.

>
>> }

2017-12-12 18:19:54

by James Morse

Subject: Re: [RESEND PATCH V2] arm64: fault: avoid send SIGBUS two times

Hi Dongjiu Geng,

On 11/12/17 16:05, Dongjiu Geng wrote:
> do_sea() calls arm64_notify_die() which will always signal
> user-space. It also returns whether APEI claimed the external
> abort as a RAS notification. If it returns failure do_mem_abort()
> will signal user-space too.
>
> do_mem_abort() wants to know if we handled the error, we always
> call arm64_notify_die() so can always return success.
>
> Signed-off-by: Dongjiu Geng <[email protected]>

Reviewed-by: James Morse <[email protected]>


Nit: Your 'RESEND V2' and 'V2' are not the same patch.
'RESEND' is to indicate you're reposting exactly the same patch, usually with a
fixed CC list. Anyone who receives both can ignore one as you've said they are
the same.



Thanks,

James

2017-12-13 02:17:28

by Dongjiu Geng

Subject: Re: [RESEND PATCH V2] arm64: fault: avoid send SIGBUS two times

> Reviewed-by: James Morse <[email protected]>
>
>
> Nit: Your 'RESEND V2' and 'V2' are not the same patch.
> 'RESEND' is to indicate you're reposting exactly the same patch, usually with a
> fixed CC list. Anyone who receives both can ignore one as you've said they are
> the same.

James,
Thanks for the reminder and guild, I will send a version V3 to eliminate this confusion