This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFC PATCH 0/3] bugfix for aarch64 ILP32


On 2015/3/18 19:38, Szabolcs Nagy wrote:
> 
> 
> On 18/03/15 10:30, Zhang Jian(Bamvor) wrote:
>> This is bamvor from Huawei OS team. We test ILP32 support through
>> LTP testcases for both little endian and big endian. There are some
>> failures we thought they should be fixed in glibc level, so, here they are.
>>
>> Hope these series of patch could help other people who is interested
>> in ILP32 in aarch64. We will send another series in LTP mailing list about
>> LTP relative patches.
>>
>> These work is base on Andrew's ilp32 patch [ilp32 v3 patch]
>> [ilp32_glibc_patch].
>>
>> The test of build is based on the toolchain built by linaro ABE environment
>> [linaro_abe].
>>
>> [ilp32 v3 patch] (https://lkml.org/lkml/2014/9/3/704)
>> [ilp32_glibc_patch] (https://sourceware.org/ml/libc-alpha/2014-10/msg00596.html)
>> [linaro_abe] (https://wiki.linaro.org/ABE)
> 
> (added Marcus to cc)
> 
> i hope you followed the recent discussion about ilp32 kernel uapi:
> 
> https://lkml.org/lkml/2015/2/10/433
> 
> in short some of the conformance issues should be fixed on the
> kernel side, so the current ilp32 patch is not ready for upstream
Yes, I think I follow the short version above. I am not try if missing
something in above long discussion. Anyway, I will read it again.
>
>(please share your opinion there if you have some particular
>ilp32 usecase in mind that depends on kernel uapi behaviour)
There was a patch about change kernel uapi to as same as the arm 32bit
version[1]. That patch try to keep the userspace compatabilities. What
do you think about it?
Besides, there is a off_t relative fix in LTP testsuite[2] which is
fix sizeof(off_t) changes in application. And use 64bit off_t could
make use of the normal syscall instead of the compat one. It seems that
it worth to change.
>
>> Yang Yingliang (3):
>>   ARM64: ILP32: change register x1 to PTR_REG
>>   ARM64: ILP32: use __fsword_t in generic/bits/statfs.h
>>   ARM64: ILP32: change register x##R to PTR_REG (R)
>>
>
>the PTR_REG changes look ok to me, but depend on the aarch64 ilp32 patches
Understand, I want to do something to see if we could speed up ilp32
upstreaming.

>
>the __fsword_t change looks ok too, not sure how statfs worked for x32 before
Yeah, I will try the H,J Lu's patch.

regards

bamvor
>
>>  sysdeps/aarch64/sysdep.h                      |  2 +-
>>  sysdeps/unix/sysv/linux/aarch64/sysdep.h      |  2 +-
>>  sysdeps/unix/sysv/linux/generic/bits/statfs.h | 24 ++++++++++++------------
>>  3 files changed, 14 insertions(+), 14 deletions(-)
>>
>
>.
[1] https://sourceware.org/ml/libc-alpha/2015-03/msg00579.html
[2] http://sourceforge.net/p/ltp/mailman/ltp-list/thread/1426742198-17967-3-git-send-email-bamvor.zhangjian%40huawei.com/#msg33613776



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]