This is the mail archive of the libc-ports@sources.redhat.com mailing list for the libc-ports 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: [PATCH] arm: ioperm use /proc/sys not sysctl


"Carlos O'Donell" <carlos@systemhalted.org> writes:

> On Mon, Nov 9, 2009 at 12:58 AM, Eric W. Biederman
> <ebiederm@xmission.com> wrote:
>>
>> sysctl is essentially unmaintained bloat in the linux kernel
>> and is scheduled for removal in a year or so, meanwhile the
>> /proc/sys interface will remain and has always existed.
>>
>> So use /proc/sys instead of sysctl in the implementation
>> of ioperm so it continues to work even when sysctl is not
>> present in the kernel.
>>
>> I don't have an arm build system setup so I can not easily
>> verify I haven't made a stupid typo. Âget_sysctl_int does
>> with other sysctl files on my x86 test machine.
>
> Please review the contribution checklist:
> http://sourceware.org/glibc/wiki/Contribution%20checklist

That checklist appears to be mindless foolish bureaucracy that is
mind numbing to read, and it was impossible to find before I tried
to help.

> It is very important to test your changes on the target machine. Keep
> in mind that in many cases testing is the difficult work.

I tested it to the extent that I am certain it will work if it happens
to compile in glibc weird environment.

> At the very least your need:
> * A ChangeLog

I had one.  Not the glibc format but I had one.

> * Testing on ARM.
> * Proof of FSF copyright assignment.

I am under 15 lines and no the FSF may not have my copyright.  I'm
certainly not going through all of that rigmarole for an issue that
should have been reported every time some uses ioperm/inb/outb on
arm.

Are there people here who actually care about glibc functioning well?

My apologies if this comes off rude.  But I am rather insulted by a
response that seems to be saying I am not good enough to talk to you,
and you won't even take your patch as a starting place for a
discussion on how to make something work properly.

I sent the patch because that function is the only significant user of
sys_sysctl on linux.  This reception makes me regret I sent the patch.
If a blow off is all I am going to get, well sysctl goes away in
September 2010 and I wish you well with the pieces.

Eric


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