Message-ID: <1875ab7a-204e-4150-c7cc-d282f69da724@arm.com>
Date:   Tue, 11 Jun 2019 10:45:30 +0530
From:   Anshuman Khandual <anshuman.khandual@....com>
To:     Christophe Leroy <christophe.leroy@....fr>,
        linux-kernel@...r.kernel.org, linux-mm@...ck.org
Cc:     linux-arm-kernel@...ts.infradead.org, linux-ia64@...r.kernel.org,
        linuxppc-dev@...ts.ozlabs.org, linux-s390@...r.kernel.org,
        linux-sh@...r.kernel.org, sparclinux@...r.kernel.org,
        x86@...nel.org, Andrew Morton <akpm@...ux-foundation.org>,
        Michal Hocko <mhocko@...e.com>,
        Matthew Wilcox <willy@...radead.org>,
        Mark Rutland <mark.rutland@....com>,
        Stephen Rothwell <sfr@...b.auug.org.au>,
        Andrey Konovalov <andreyknvl@...gle.com>,
        Michael Ellerman <mpe@...erman.id.au>,
        Paul Mackerras <paulus@...ba.org>,
        Russell King <linux@...linux.org.uk>,
        Catalin Marinas <catalin.marinas@....com>,
        Will Deacon <will.deacon@....com>,
        Tony Luck <tony.luck@...el.com>,
        Fenghua Yu <fenghua.yu@...el.com>,
        Martin Schwidefsky <schwidefsky@...ibm.com>,
        Heiko Carstens <heiko.carstens@...ibm.com>,
        Yoshinori Sato <ysato@...rs.sourceforge.jp>,
        "David S. Miller" <davem@...emloft.net>,
        Thomas Gleixner <tglx@...utronix.de>,
        Peter Zijlstra <peterz@...radead.org>,
        Ingo Molnar <mingo@...hat.com>,
        Andy Lutomirski <luto@...nel.org>,
        Dave Hansen <dave.hansen@...ux.intel.com>
Subject: Re: [RFC V3] mm: Generalize and rename notify_page_fault() as
 kprobe_page_fault()



On 06/11/2019 10:16 AM, Christophe Leroy wrote:
> 
> 
> Le 10/06/2019 à 04:39, Anshuman Khandual a écrit :
>>
>>
>> On 06/07/2019 09:01 PM, Christophe Leroy wrote:
>>>
>>>
>>> Le 07/06/2019 à 12:34, Anshuman Khandual a écrit :
>>>> Very similar definitions for notify_page_fault() are being used by multiple
>>>> architectures duplicating much of the same code. This attempts to unify all
>>>> of them into a generic implementation, rename it as kprobe_page_fault() and
>>>> then move it to a common header.
>>>>
>>>> kprobes_built_in() can detect CONFIG_KPROBES, hence new kprobe_page_fault()
>>>> need not be wrapped again within CONFIG_KPROBES. Trap number argument can
>>>> now contain upto an 'unsigned int' accommodating all possible platforms.
>>>>
>>>> kprobe_page_fault() goes the x86 way while dealing with preemption context.
>>>> As explained in these following commits the invoking context in itself must
>>>> be non-preemptible for kprobes processing context irrespective of whether
>>>> kprobe_running() or perhaps smp_processor_id() is safe or not. It does not
>>>> make much sense to continue when original context is preemptible. Instead
>>>> just bail out earlier.
>>>>
>>>> commit a980c0ef9f6d
>>>> ("x86/kprobes: Refactor kprobes_fault() like kprobe_exceptions_notify()")
>>>>
>>>> commit b506a9d08bae ("x86: code clarification patch to Kprobes arch code")
>>>>
>>>> Cc: linux-arm-kernel@...ts.infradead.org
>>>> Cc: linux-ia64@...r.kernel.org
>>>> Cc: linuxppc-dev@...ts.ozlabs.org
>>>> Cc: linux-s390@...r.kernel.org
>>>> Cc: linux-sh@...r.kernel.org
>>>> Cc: sparclinux@...r.kernel.org
>>>> Cc: x86@...nel.org
>>>> Cc: Andrew Morton <akpm@...ux-foundation.org>
>>>> Cc: Michal Hocko <mhocko@...e.com>
>>>> Cc: Matthew Wilcox <willy@...radead.org>
>>>> Cc: Mark Rutland <mark.rutland@....com>
>>>> Cc: Christophe Leroy <christophe.leroy@....fr>
>>>> Cc: Stephen Rothwell <sfr@...b.auug.org.au>
>>>> Cc: Andrey Konovalov <andreyknvl@...gle.com>
>>>> Cc: Michael Ellerman <mpe@...erman.id.au>
>>>> Cc: Paul Mackerras <paulus@...ba.org>
>>>> Cc: Russell King <linux@...linux.org.uk>
>>>> Cc: Catalin Marinas <catalin.marinas@....com>
>>>> Cc: Will Deacon <will.deacon@....com>
>>>> Cc: Tony Luck <tony.luck@...el.com>
>>>> Cc: Fenghua Yu <fenghua.yu@...el.com>
>>>> Cc: Martin Schwidefsky <schwidefsky@...ibm.com>
>>>> Cc: Heiko Carstens <heiko.carstens@...ibm.com>
>>>> Cc: Yoshinori Sato <ysato@...rs.sourceforge.jp>
>>>> Cc: "David S. Miller" <davem@...emloft.net>
>>>> Cc: Thomas Gleixner <tglx@...utronix.de>
>>>> Cc: Peter Zijlstra <peterz@...radead.org>
>>>> Cc: Ingo Molnar <mingo@...hat.com>
>>>> Cc: Andy Lutomirski <luto@...nel.org>
>>>> Cc: Dave Hansen <dave.hansen@...ux.intel.com>
>>>>
>>>> Signed-off-by: Anshuman Khandual <anshuman.khandual@....com>
>>>> ---
>>>> Testing:
>>>>
>>>> - Build and boot tested on arm64 and x86
>>>> - Build tested on some other archs (arm, sparc64, alpha, powerpc etc)
>>>>
>>>> Changes in RFC V3:
>>>>
>>>> - Updated the commit message with an explaination for new preemption behaviour
>>>> - Moved notify_page_fault() to kprobes.h with 'static nokprobe_inline' per Matthew
>>>> - Changed notify_page_fault() return type from int to bool per Michael Ellerman
>>>> - Renamed notify_page_fault() as kprobe_page_fault() per Peterz
>>>>
>>>> Changes in RFC V2: (https://patchwork.kernel.org/patch/10974221/)
>>>>
>>>> - Changed generic notify_page_fault() per Mathew Wilcox
>>>> - Changed x86 to use new generic notify_page_fault()
>>>> - s/must not/need not/ in commit message per Matthew Wilcox
>>>>
>>>> Changes in RFC V1: (https://patchwork.kernel.org/patch/10968273/)
>>>>
>>>>    arch/arm/mm/fault.c      | 24 +-----------------------
>>>>    arch/arm64/mm/fault.c    | 24 +-----------------------
>>>>    arch/ia64/mm/fault.c     | 24 +-----------------------
>>>>    arch/powerpc/mm/fault.c  | 23 ++---------------------
>>>>    arch/s390/mm/fault.c     | 16 +---------------
>>>>    arch/sh/mm/fault.c       | 18 ++----------------
>>>>    arch/sparc/mm/fault_64.c | 16 +---------------
>>>>    arch/x86/mm/fault.c      | 21 ++-------------------
>>>>    include/linux/kprobes.h  | 16 ++++++++++++++++
>>>>    9 files changed, 27 insertions(+), 155 deletions(-)
>>>>
>>>
>>> [...]
>>>
>>>> diff --git a/include/linux/kprobes.h b/include/linux/kprobes.h
>>>> index 443d980..064dd15 100644
>>>> --- a/include/linux/kprobes.h
>>>> +++ b/include/linux/kprobes.h
>>>> @@ -458,4 +458,20 @@ static inline bool is_kprobe_optinsn_slot(unsigned long addr)
>>>>    }
>>>>    #endif
>>>>    +static nokprobe_inline bool kprobe_page_fault(struct pt_regs *regs,
>>>> +                          unsigned int trap)
>>>> +{
>>>> +    int ret = 0;
>>>
>>> ret is pointless.
>>>
>>>> +
>>>> +    /*
>>>> +     * To be potentially processing a kprobe fault and to be allowed
>>>> +     * to call kprobe_running(), we have to be non-preemptible.
>>>> +     */
>>>> +    if (kprobes_built_in() && !preemptible() && !user_mode(regs)) {
>>>> +        if (kprobe_running() && kprobe_fault_handler(regs, trap))
>>>
>>> don't need an 'if A if B', can do 'if A && B'
>>
>> Which will make it a very lengthy condition check.
> 
> Yes. But is that a problem at all ?

Probably not.

> 
> For me the following would be easier to read.
> 
> if (kprobes_built_in() && !preemptible() && !user_mode(regs) &&
>     kprobe_running() && kprobe_fault_handler(regs, trap))
>     ret = 1;

As mentioned before will stick with current x86 implementation. 

Powered by blists - more mailing lists