[prev in list] [next in list] [prev in thread] [next in thread] 

List:       qemu-ppc
Subject:    Re: [Qemu-ppc] [PATCH] linux-user: elf: mmap all the target-pages of hostpage for data segment
From:       Shivaprasad G Bhat <sbhat () linux ! vnet ! ibm ! com>
Date:       2018-08-28 15:35:42
Message-ID: 901aa049-0acf-0f6b-4257-9d785713d2cd () linux ! vnet ! ibm ! com
[Download RAW message or body]



On 08/27/2018 06:55 PM, Laurent Vivier wrote:
> Le 27/08/2018 à 14:37, Shivaprasad G Bhat a écrit  :
>> If the hostpage size is greater than the TARGET_PAGESIZE, the
>> target-pages of size TARGET_PAGESIZE are marked valid only till the
>> length requested during the elfload. The glibc attempts to consume unused
>> space in the last page of data segment(__libc_memalign() in
>> elf/dl-minimal.c). The GLRO(dl_pagesize) is actually the host pagesize as
>> set in the auxillary vectors. So, there is no explicit mmap request for
>> the remaining target-pages on the last hostpage. The glibc assumes that
>> particular space as available and subsequent attempts to use
>> those addresses lead to crash as the target_mmap has not marked them valid
>> for those target-pages.
>>
>> The issue is seen when trying to chroot to 16.04-x86_64 ubuntu on a PPC64
>> host where the fork fails to access the thread_id as it is allocated on a
>> page not marked valid. The recent glibc doesnt have checks for thread-id in
>> fork, but the issue can manifest somewhere else, none the less.
>>
>> The fix here is to map all the target-pages of the hostpage during the
>> ELF load for data segment to allow the glibc for proper consumption.
>>
>> Signed-off-by: Shivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
>> ---
>>   linux-user/elfload.c |   24 +++++++++++++++++-------
>>   1 file changed, 17 insertions(+), 7 deletions(-)
>>
>> diff --git a/linux-user/elfload.c b/linux-user/elfload.c
>> index 8638612aec..1d86034c8d 100644
>> --- a/linux-user/elfload.c
>> +++ b/linux-user/elfload.c
>> @@ -1438,9 +1438,17 @@ struct exec
>>   
>>   /* Necessary parameters */
>>   #define TARGET_ELF_EXEC_PAGESIZE TARGET_PAGE_SIZE
>> -#define TARGET_ELF_PAGESTART(_v) ((_v) & \
>> -                                 ~(abi_ulong)(TARGET_ELF_EXEC_PAGESIZE-1))
>> -#define TARGET_ELF_PAGEOFFSET(_v) ((_v) & (TARGET_ELF_EXEC_PAGESIZE-1))
>> +#define TARGET_ELF_PAGESTART(_v, _s) \
>> +        ((TARGET_ELF_EXEC_PAGESIZE > _s) ? \
>> +         (_v) & ~(abi_ulong)(TARGET_ELF_EXEC_PAGESIZE - 1) : \
>> +         (_v) & ~(abi_ulong)(_s - 1));
>> +#define TARGET_ELF_PAGEOFFSET(_v, _s) \
>> +        ((TARGET_ELF_EXEC_PAGESIZE > _s) ? \
>> +         (_v) & (TARGET_ELF_EXEC_PAGESIZE - 1) : \
>> +         (_v) & (_s - 1));
>> +#define TARGET_ELF_PAGELENGTH(_v, _s) \
>> +        ((TARGET_ELF_EXEC_PAGESIZE > _s) ? \
>> +         TARGET_PAGE_ALIGN(_v) : HOST_PAGE_ALIGN(_v));
> I think it's only possible if the PT_LOAD p_align value is greater or
> equal to qemu_host_page_size.
>
> See 33143c446e ("linux-user: fix ELF load alignment error").
>
> You could check this with qemu-s390x or qemu-arm on ppc64 host.
Ah, right! I should have added the extra conditional to check the 
p_align with qemu_host_page_mask
along with the existing ones.

Posted the v2 accordingly.

Thanks and Regards,
Shivaprasad

> Thanks,
> Laurent
>



[prev in list] [next in list] [prev in thread] [next in thread]