On Mon, Oct 25, 2010 at 11:11 PM, Khem Raj <[email protected]> wrote:
> On (18/10/10 15:09), Natanael Copa wrote:
>> I (or an Alpine linux user) got bitten by the issue below too.
>> http://redmine.alpinelinux.org/issues/show/445
>>
>> So I started to dig around a bit. Here is what i have come up with:
>> * it happens with busybox ls and coreutils ls.
>> * it works on 32bit (e)glibc systems. ie ubuntu.
>>
>> So I compared the straces and as expected, the difference is stat()
>> and stat64().
>>
>> >From uclibc:
>> open("/mnt/iso/", O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = 3
>> fstat(3, 0xbffff148)                    = -1 EOVERFLOW (Value too
>> large for defined data type)
>>
>> >From ubuntu/eglibc:
>> open("/mnt/iso/", O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY|0x80000) = 3
>> fstat64(3, {st_mode=S_IFDIR|0777, st_size=4096, ...}) = 0
>>
>> Then i thought that there are probably some flag that is supposed to
>> do so fstat = fstat64 or something, and if i know those GNU people
>> correctly, they check this in a configure script. this turned out to
>> be correct. In the configure output I see this when compiling
>> coreutils:
>>
>> checking for _LARGEFILE_SOURCE value needed for large files... no
>> ...
>> checking for _FILE_OFFSET_BITS value needed for large files... 64
>> ...
>>
>> So, I think uclibc is not doing what expected at some place.
>> Where or how, I don't know.

>
> Do you have large file support turned on in uclibc config ?

I have.

-- 
Natanael Copa
_______________________________________________
uClibc mailing list
[email protected]
http://lists.busybox.net/mailman/listinfo/uclibc

Reply via email to