[PATCH] seq_file:update file->f_pos when lseek() to m->read_pos

[PATCH] seq_file:update file->f_pos when lseek() to m->read_pos

[Jiaxing Wang]

After pread(), file->f_pos and m->read_pos get different,
and lseek() to m->read_pos did not update file->f_pos, then
a subsequent read may read from a wrong position, the following
program shows the problem:

    char str1[32] = { 0 };
    char str2[32] = { 0 };
    int poffset = 10;
    int count = 20;

    /*open any seq file*/
    int fd = open("/proc/modules", O_RDONLY);

    pread(fd, str1, count, poffset);
    printf("pread:%s\n", str1);

    /*seek to where m->read_pos is*/
    lseek(fd, poffset+count, SEEK_SET);

    /*supposed to read from poffset+count, but this read from position 0*/
    read(fd, str2, count);
    printf("read:%s\n", str2);

Signed-off-by: Jiaxing Wang <hello.wjx@gmail.com>
---
 fs/seq_file.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/fs/seq_file.c b/fs/seq_file.c
index 774c1eb..4b22b26 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -328,7 +328,8 @@ loff_t seq_lseek(struct file *file, loff_t offset, int whence)
 				m->read_pos = offset;
 				retval = file->f_pos = offset;
 			}
-		}
+		} else
+			file->f_pos = offset;
 	}
 	file->f_version = m->version;
 	mutex_unlock(&m->lock);
-- 
1.8.1.2

Re: [PATCH] seq_file:update file->f_pos when lseek() to m->read_pos

[fangdong]

On 06/29/2013 05:11 AM, Jiaxing Wang wrote:
> After pread(), file->f_pos and m->read_pos get different,
> and lseek() to m->read_pos did not update file->f_pos, then
> a subsequent read may read from a wrong position, the following
> program shows the problem:
>
>      char str1[32] = { 0 };
>      char str2[32] = { 0 };
>      int poffset = 10;
>      int count = 20;
>
>      /*open any seq file*/
>      int fd = open("/proc/modules", O_RDONLY);
>
>      pread(fd, str1, count, poffset);
>      printf("pread:%s\n", str1);
>
>      /*seek to where m->read_pos is*/
>      lseek(fd, poffset+count, SEEK_SET);
>
>      /*supposed to read from poffset+count, but this read from position 0*/
>      read(fd, str2, count);
>      printf("read:%s\n", str2);
>
> Signed-off-by: Jiaxing Wang <hello.wjx@gmail.com>
> ---
>   fs/seq_file.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/fs/seq_file.c b/fs/seq_file.c
> index 774c1eb..4b22b26 100644
> --- a/fs/seq_file.c
> +++ b/fs/seq_file.c
> @@ -328,7 +328,8 @@ loff_t seq_lseek(struct file *file, loff_t offset, int whence)
>   				m->read_pos = offset;
>   				retval = file->f_pos = offset;
>   			}
> -		}
> +		} else
> +			file->f_pos = offset;
>   	}
>   	file->f_version = m->version;
>   	mutex_unlock(&m->lock);
>
This does not appear to be a problem, in linux man page, the behaver 
seems clearly defined:

DESCRIPTION
        pread() reads up to count bytes from file descriptor fd at 
offset  off-set  (from the start of the file) into the buffer starting 
at buf.  The file offset is not changed.

        pwrite() writes up to count bytes from the buffer starting at 
buf  to the  file  descriptor  fd  at  offset  offset.   The file offset 
is not changed.

Re: [PATCH] seq_file:update file->f_pos when lseek() to m->read_pos

[Dong Fang]

On 06/29/2013 05:11 AM, Jiaxing Wang wrote:
> After pread(), file->f_pos and m->read_pos get different,
> and lseek() to m->read_pos did not update file->f_pos, then
> a subsequent read may read from a wrong position, the following
> program shows the problem:
>
>      char str1[32] = { 0 };
>      char str2[32] = { 0 };
>      int poffset = 10;
>      int count = 20;
>
>      /*open any seq file*/
>      int fd = open("/proc/modules", O_RDONLY);
>
>      pread(fd, str1, count, poffset);
>      printf("pread:%s\n", str1);
>
>      /*seek to where m->read_pos is*/
>      lseek(fd, poffset+count, SEEK_SET);
>
>      /*supposed to read from poffset+count, but this read from position 0*/
>      read(fd, str2, count);
>      printf("read:%s\n", str2);
>
> Signed-off-by: Jiaxing Wang <hello.wjx@gmail.com>
> ---
>   fs/seq_file.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/fs/seq_file.c b/fs/seq_file.c
> index 774c1eb..4b22b26 100644
> --- a/fs/seq_file.c
> +++ b/fs/seq_file.c
> @@ -328,7 +328,8 @@ loff_t seq_lseek(struct file *file, loff_t offset, int whence)
>   				m->read_pos = offset;
>   				retval = file->f_pos = offset;
>   			}
> -		}
> +		} else
> +			file->f_pos = offset;
>   	}
>   	file->f_version = m->version;
>   	mutex_unlock(&m->lock);
>
This does not appear to be a problem, in linux man page, the behaver 
seems clearly defined:

DESCRIPTION
        pread() reads up to count bytes from file descriptor fd at
offset  off-set  (from the start of the file) into the buffer starting
at buf.  The file offset is not changed.
        pwrite() writes up to count bytes from the buffer starting at
buf  to the  file  descriptor  fd  at  offset  offset.   The file
offset is not changed.

Re: [PATCH] seq_file:update file->f_pos when lseek() to m->read_pos

[Jiaxing Wang]

On 07/01/2013 08:41 PM, fangdong wrote:
> On 06/29/2013 05:11 AM, Jiaxing Wang wrote:
>> After pread(), file->f_pos and m->read_pos get different,
>> and lseek() to m->read_pos did not update file->f_pos, then
>> a subsequent read may read from a wrong position, the following
>> program shows the problem:
>>
>>      char str1[32] = { 0 };
>>      char str2[32] = { 0 };
>>      int poffset = 10;
>>      int count = 20;
>>
>>      /*open any seq file*/
>>      int fd = open("/proc/modules", O_RDONLY);
>>
>>      pread(fd, str1, count, poffset);
>>      printf("pread:%s\n", str1);
>>
>>      /*seek to where m->read_pos is*/
>>      lseek(fd, poffset+count, SEEK_SET);
>>
>>      /*supposed to read from poffset+count, but this read from position 0*/
>>      read(fd, str2, count);
>>      printf("read:%s\n", str2);
>>
>> Signed-off-by: Jiaxing Wang <hello.wjx@gmail.com>
>> ---
>>   fs/seq_file.c | 3 ++-
>>   1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/fs/seq_file.c b/fs/seq_file.c
>> index 774c1eb..4b22b26 100644
>> --- a/fs/seq_file.c
>> +++ b/fs/seq_file.c
>> @@ -328,7 +328,8 @@ loff_t seq_lseek(struct file *file, loff_t offset, int whence)
>>                   m->read_pos = offset;
>>                   retval = file->f_pos = offset;
>>               }
>> -        }
>> +        } else
>> +            file->f_pos = offset;
>>       }
>>       file->f_version = m->version;
>>       mutex_unlock(&m->lock);
>>
> This does not appear to be a problem, in linux man page, the behaver seems clearly defined:
>
> DESCRIPTION
>        pread() reads up to count bytes from file descriptor fd at offset  off-set  (from the start of the file) into the buffer starting at buf.  The file offset is not changed.
>
>        pwrite() writes up to count bytes from the buffer starting at buf  to the  file  descriptor  fd  at  offset  offset.   The file offset is not changed.
>
> -- 
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
There's no problem that pread() don't change file->f_pos, but I think lseek() should have changed it.
Any comments from other people, and Alexander?

Re: [PATCH] seq_file:update file->f_pos when lseek() to m->read_pos

[Gu Zheng]

Hi Jiaxing,
	Please refer to inline comment.:)
On 07/02/2013 08:43 AM, Jiaxing Wang wrote:

> On 07/01/2013 08:41 PM, fangdong wrote:
>> On 06/29/2013 05:11 AM, Jiaxing Wang wrote:
>>> After pread(), file->f_pos and m->read_pos get different,
>>> and lseek() to m->read_pos did not update file->f_pos, then
>>> a subsequent read may read from a wrong position, the following
>>> program shows the problem:
>>>
>>>      char str1[32] = { 0 };
>>>      char str2[32] = { 0 };
>>>      int poffset = 10;
>>>      int count = 20;
>>>
>>>      /*open any seq file*/
>>>      int fd = open("/proc/modules", O_RDONLY);
>>>
>>>      pread(fd, str1, count, poffset);
>>>      printf("pread:%s\n", str1);
>>>
>>>      /*seek to where m->read_pos is*/
>>>      lseek(fd, poffset+count, SEEK_SET);
>>>
>>>      /*supposed to read from poffset+count, but this read from position 0*/
>>>      read(fd, str2, count);
>>>      printf("read:%s\n", str2);
>>>
>>> Signed-off-by: Jiaxing Wang <hello.wjx@gmail.com>
>>> ---
>>>   fs/seq_file.c | 3 ++-
>>>   1 file changed, 2 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/fs/seq_file.c b/fs/seq_file.c
>>> index 774c1eb..4b22b26 100644
>>> --- a/fs/seq_file.c
>>> +++ b/fs/seq_file.c
>>> @@ -328,7 +328,8 @@ loff_t seq_lseek(struct file *file, loff_t offset, int whence)
>>>                   m->read_pos = offset;
>>>                   retval = file->f_pos = offset;
>>>               }
>>> -        }
>>> +        } else
>>> +            file->f_pos = offset;
>>>       }
>>>       file->f_version = m->version;
>>>       mutex_unlock(&m->lock);
>>>
>> This does not appear to be a problem, in linux man page, the behaver seems clearly defined:
>>
>> DESCRIPTION
>>        pread() reads up to count bytes from file descriptor fd at offset  off-set  (from the start of the file) into the buffer starting at buf.  The file offset is not changed.
>>
>>        pwrite() writes up to count bytes from the buffer starting at buf  to the  file  descriptor  fd  at  offset  offset.   The file offset is not changed.
>>
>> -- 
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at  http://www.tux.org/lkml/
> There's no problem that pread() don't change file->f_pos, but I think lseek() should have changed it.

No, the way that seq_seek deal with pos is right. Remember this seq file, it is
an iterator interface, e.g. it provides infos of each element of a list, so the
we should make sure that the info we read from seq file is the whole infos of
each entry, so seq_lseek can not set pos to a optional point, only can on the
points that divided by a whole infos size of each entry. If we can not set pos
on the special position, ste it as 0, I think it's reasonable.

Thanks,
Gu  

> Any comments from other people, and Alexander?
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>