Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <6896bb4c-5224-678f-e73c-fe95be41d196@loongson.cn>
Date:   Thu, 14 Jul 2022 22:07:52 +0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.11.0
Subject: Re: [PATCH] LoongArch: Fix missing fcsr in ptrace's fpr_set
Content-Language: en-US
To:     Huacai Chen <chenhuacai@kernel.org>
Cc:     WANG Xuerui <kernel@xen0n.name>, Oleg Nesterov <oleg@redhat.com>,
        Xu Li <lixu@loongson.cn>, loongarch@lists.linux.dev,
        LKML <linux-kernel@vger.kernel.org>
References: <20220714062550.4934-1-huqi@loongson.cn>
 <CAAhV-H6NWYG1ChLRyK-EAGAmhDvP-x+z9BgBo3a=GAbQ-2bKig@mail.gmail.com>
From:   Qi Hu <huqi@loongson.cn>
In-Reply-To: <CAAhV-H6NWYG1ChLRyK-EAGAmhDvP-x+z9BgBo3a=GAbQ-2bKig@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 2022/7/14 21:12, Huacai Chen wrote:
> Hi,
>
> On Thu, Jul 14, 2022 at 2:26 PM Qi Hu <huqi@loongson.cn> wrote:
>> In file ptrace.c, function fpr_set does not copy fcsr data from ubuf
>> to kbuf. That's the reason why fcsr cannot be modified by ptrace.
>>
>> This patch fixs this problem and allows users using ptrace to modify
>> the fcsr.
>>
>> Signed-off-by: Qi Hu <huqi@loongson.cn>
>> Signed-off-by: Xu Li <lixu@loongson.cn>
>> ---
>>   arch/loongarch/kernel/ptrace.c | 12 +++++++-----
>>   1 file changed, 7 insertions(+), 5 deletions(-)
>>
>> diff --git a/arch/loongarch/kernel/ptrace.c b/arch/loongarch/kernel/ptrace.c
>> index e6ab87948e1d..dc2b82ea894c 100644
>> --- a/arch/loongarch/kernel/ptrace.c
>> +++ b/arch/loongarch/kernel/ptrace.c
>> @@ -193,7 +193,7 @@ static int fpr_set(struct task_struct *target,
>>                     const void *kbuf, const void __user *ubuf)
>>   {
>>          const int fcc_start = NUM_FPU_REGS * sizeof(elf_fpreg_t);
>> -       const int fcc_end = fcc_start + sizeof(u64);
>> +       const int fcsr_start = fcc_start + sizeof(u64);
>>          int err;
>>
>>          BUG_ON(count % sizeof(elf_fpreg_t));
>> @@ -209,10 +209,12 @@ static int fpr_set(struct task_struct *target,
>>          if (err)
>>                  return err;
>>
>> -       if (count > 0)
>> -               err |= user_regset_copyin(&pos, &count, &kbuf, &ubuf,
>> -                                         &target->thread.fpu.fcc,
>> -                                         fcc_start, fcc_end);
>> +       err |= user_regset_copyin(&pos, &count, &kbuf, &ubuf,
>> +                                 &target->thread.fpu.fcc, fcc_start,
>> +                                 fcc_start + sizeof(u64));
>> +       err |= user_regset_copyin(&pos, &count, &kbuf, &ubuf,
>> +                                 &target->thread.fpu.fcsr, fcsr_start,
>> +                                 fcsr_start + sizeof(u32));
> You shouldn't remove (count > 0) here, because  the above
> user_regset_copyin() will modify count inside, and so "count == 0" is
> possible.
>
> Huacai

Yes, the "count" should be considered. But the "count" is checked at the 
beginning of the "user_regset_copyin()".

So "count > 0" is useless, i think.

Also like riscv, "count" is not checked in "riscv_fpr_set()".

>>          return err;
>>   }
>> --
>> 2.37.0
>>