[PATCH v4 7/8] vmcore: treat memory chunks referenced by PT_LOAD program header entries in page-size boundary in vmcore_list

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Sorry for late reply.

(2013/04/30 4:51), Vivek Goyal wrote:
> On Sat, Apr 13, 2013 at 09:21:46AM +0900, HATAYAMA Daisuke wrote:
>> Treat memory chunks referenced by PT_LOAD program header entries in
>> page-size boundary in vmcore_list. Formally, for each range [start,
>> end], we set up the corresponding vmcore object in vmcore_list to
>> [rounddown(start, PAGE_SIZE), roundup(end, PAGE_SIZE)].
>>
>> This change affects layout of /proc/vmcore. The gaps generated by the
>> rearrangement are newly made visible to applications as
>> holes. Concretely, they are two ranges [rounddown(start, PAGE_SIZE),
>> start] and [end, roundup(end, PAGE_SIZE)].
>
> Sorry did not understand this part. So if end is not page aligned, then
> we roundup(end, PAGE_SIZE) and increase the PT_LOAD size accordingly?
> Similarly for start, we do roundown().
>
> - Can you really rounddown() start? Then you will have to change start
>    virtual address in program header and that's not really a good idea.
>

No, there's no need to change paddr in program header. Pleaes notice 
that difference between what objects in vc_list refer to and what 
PT_LOAD program headers refer to. The former covers not only kernel 
memory but also the extra memory, while the latter the kernel memory only.

> - So this extra memory for end, we read from old memory and not fill
>    with zeros?

Yes. The extra memory is not covered by any program header, i.e. hole. 
The extra memory is not modified at all, exported with no change; if it 
is used by BIOS, users can see BIOS data there. This design comes from 
the discussion with Erick.

>
>>
>> Signed-off-by: HATAYAMA Daisuke <d.hatayama at jp.fujitsu.com>
>> ---
>>
>>   fs/proc/vmcore.c |   26 ++++++++++++++++++++------
>>   1 files changed, 20 insertions(+), 6 deletions(-)
>>
>> diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
>> index 029bdc0..cd0f9d9 100644
>> --- a/fs/proc/vmcore.c
>> +++ b/fs/proc/vmcore.c
>> @@ -477,16 +477,23 @@ static int __init process_ptload_program_headers_elf64(char *elfptr,
>>   	vmcore_off = elfsz + roundup(phdr_ptr->p_memsz, PAGE_SIZE);
>>
>>   	for (i = 0; i < ehdr_ptr->e_phnum; i++, phdr_ptr++) {
>> +		u64 paddr, start, end, size;
>> +
>>   		if (phdr_ptr->p_type != PT_LOAD)
>>   			continue;
>>
>> +		paddr = phdr_ptr->p_offset;
>> +		start = rounddown(paddr, PAGE_SIZE);
>> +		end = roundup(paddr + phdr_ptr->p_memsz, PAGE_SIZE);
>> +		size = end - start;
>> +
>>   		/* Add this contiguous chunk of memory to vmcore list.*/
>> -		if (vmcore_add(vc_list, phdr_ptr->p_offset, phdr_ptr->p_memsz))
>> +		if (vmcore_add(vc_list, start, size))
>>   			return -ENOMEM;
>>
>>   		/* Update the program header offset. */
>> -		phdr_ptr->p_offset = vmcore_off;
>> -		vmcore_off = vmcore_off + phdr_ptr->p_memsz;
>> +		phdr_ptr->p_offset = vmcore_off + (paddr - start);
>
> What's paddr-start. Why following is not sufficient.
>
> phdr_ptr->p_offset = vmcore_off
>

(paddr - start) is offset of the memory program header refers to, from 
which kernel memory starts. Pictrically:

   vmcore_off                   +----------------------+
                                |  extra memory        |
                                |  (non kernel memory) |
phdr->p_offset =               +----------------------+
   vmcore_off + (paddr - start) |                      |\
                                |  kernel memory       | phdr->p_memsz
                                |                      |/
                                +----------------------+
                                |  extra memory        |
                                |  (non kernel memory) |
   vmcore_off + size            +----------------------+

-- 
Thanks.
HATAYAMA, Daisuke




[Index of Archives]     [LM Sensors]     [Linux Sound]     [ALSA Users]     [ALSA Devel]     [Linux Audio Users]     [Linux Media]     [Kernel]     [Gimp]     [Yosemite News]     [Linux Media]

  Powered by Linux