Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH 10/15] powerpc/watchpoint: Use loop for
 thread_struct->ptrace_bps
To:     Christophe Leroy <christophe.leroy@c-s.fr>
Cc:     mpe@ellerman.id.au, mikey@neuling.org, apopple@linux.ibm.com,
        paulus@samba.org, npiggin@gmail.com,
        naveen.n.rao@linux.vnet.ibm.com, peterz@infradead.org,
        jolsa@kernel.org, oleg@redhat.com, fweisbec@gmail.com,
        mingo@kernel.org, linuxppc-dev@lists.ozlabs.org,
        linux-kernel@vger.kernel.org,
        Ravi Bangoria <ravi.bangoria@linux.ibm.com>
References: <20200309085806.155823-1-ravi.bangoria@linux.ibm.com>
 <20200309085806.155823-11-ravi.bangoria@linux.ibm.com>
 <0eeeac90-b5e3-722b-2d2c-bb95c81d851a@c-s.fr>
From:   Ravi Bangoria <ravi.bangoria@linux.ibm.com>
Date:   Wed, 18 Mar 2020 15:13:01 +0530
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.5.0
MIME-Version: 1.0
In-Reply-To: <0eeeac90-b5e3-722b-2d2c-bb95c81d851a@c-s.fr>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Message-Id: <a52ef8d3-b27b-d48d-b0ab-c9b41209a6ca@linux.ibm.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

>> @@ -1628,6 +1628,9 @@ int copy_thread_tls(unsigned long clone_flags, unsigned long usp,
>>       void (*f)(void);
>>       unsigned long sp = (unsigned long)task_stack_page(p) + THREAD_SIZE;
>>       struct thread_info *ti = task_thread_info(p);
>> +#ifdef CONFIG_HAVE_HW_BREAKPOINT
>> +    int i;
>> +#endif
> 
> Could we avoid all those #ifdefs ?
> 
> I think if we make p->thread.ptrace_bps[] exist all the time, with a size of 0 when CONFIG_HAVE_HW_BREAKPOINT is not set, then we can drop a lot of #ifdefs.

Hmm.. what you are saying seems possible. But IMO it should be done as
independent series. Will work on it.

> 
>>       klp_init_thread_info(p);
>> @@ -1687,7 +1690,8 @@ int copy_thread_tls(unsigned long clone_flags, unsigned long usp,
>>       p->thread.ksp_limit = (unsigned long)end_of_stack(p);
>>   #endif
>>   #ifdef CONFIG_HAVE_HW_BREAKPOINT
>> -    p->thread.ptrace_bps[0] = NULL;
>> +    for (i = 0; i < nr_wp_slots(); i++)
>> +        p->thread.ptrace_bps[i] = NULL;
>>   #endif
>>       p->thread.fp_save_area = NULL;
>> diff --git a/arch/powerpc/kernel/ptrace.c b/arch/powerpc/kernel/ptrace.c
>> index f6d7955fc61e..e2651f86d56f 100644
>> --- a/arch/powerpc/kernel/ptrace.c
>> +++ b/arch/powerpc/kernel/ptrace.c
> 
> You'll have to rebase all this on the series https://patchwork.ozlabs.org/project/linuxppc-dev/list/?series=161356 which is about to go into powerpc-next

Sure. Thanks for heads up.

> 
>> @@ -2829,6 +2829,19 @@ static int set_dac_range(struct task_struct *child,
>>   }
>>   #endif /* CONFIG_PPC_ADV_DEBUG_DAC_RANGE */
>> +#ifdef CONFIG_HAVE_HW_BREAKPOINT
>> +static int empty_ptrace_bp(struct thread_struct *thread)
>> +{
>> +    int i;
>> +
>> +    for (i = 0; i < nr_wp_slots(); i++) {
>> +        if (!thread->ptrace_bps[i])
>> +            return i;
>> +    }
>> +    return -1;
>> +}
>> +#endif
> 
> What does this function do exactly ? I seems to do more than what its name suggests.

It finds an empty breakpoint in ptrace_bps[]. But yeah, function name is
misleading. I'll rename it to find_empty_ptrace_bp().

...

>> @@ -2979,10 +2993,10 @@ static long ppc_del_hwdebug(struct task_struct *child, long data)
>>           return -EINVAL;
>>   #ifdef CONFIG_HAVE_HW_BREAKPOINT
>> -    bp = thread->ptrace_bps[0];
>> +    bp = thread->ptrace_bps[data - 1];
> 
> Is data checked somewhere to ensure it is not out of boundaries ? Or are we sure it is always within ?

Yes. it's checked. See patch #9:

   @@ -2955,7 +2975,7 @@ static long ppc_del_hwdebug(struct task_struct *child, long data)
    	}
    	return rc;
    #else
   -	if (data != 1)
   +	if (data < 1 || data > nr_wp_slots())
    		return -EINVAL;
    
    #ifdef CONFIG_HAVE_HW_BREAKPOINT

Thanks,
Ravi