1 |
On 09/10/2013 09:17 PM, Rich Freeman wrote: |
2 |
> On Tue, Sep 10, 2013 at 6:41 PM, Richard Yao <ryao@g.o> wrote: |
3 |
>> 1. The kernel expects -fno-stack-protector to be the default. What will |
4 |
>> the effect be on kernel configuration once -fstack-protector is the default? |
5 |
> |
6 |
> Nothing, since the kernel build system doesn't source make.conf. If |
7 |
> somebody creates an ebuild that actually installs a kernel then it |
8 |
> might be an issue, though it could be filtered if it is a problem. |
9 |
|
10 |
My understanding is that this change would be made to GCC's spec files, |
11 |
which affects everything that uses GCC. |
12 |
|
13 |
>> |
14 |
>> 2. We should make sure that -fno-stack-protector is a supported CFLAG. |
15 |
>> This will make it easier to handle complaints from the vocal minority of |
16 |
>> our user base that want every last percentage point of performance. |
17 |
> |
18 |
> No reason that it would be any less supported than -fstack-protector |
19 |
> already is. |
20 |
|
21 |
Just making certain. |
22 |
|
23 |
>> |
24 |
>> 3. I would like to point out that we are talking about deviating from |
25 |
>> upstream behavior and everyone is okay with it. Anyone who thinks we |
26 |
>> should stick to upstream when it is not good for us should speak now or |
27 |
>> risk being asked "where were you when..." whenever they try to use |
28 |
>> upstream as an excuse to hold back progress. ;) |
29 |
> |
30 |
> I don't really see this as an issue - in general maintainers are |
31 |
> expected to accommodate reasonable CFLAGS. This doesn't mean that |
32 |
> arbitrary CFLAGS are "supported" so much as bugs should be taken |
33 |
> seriously if they're really bugs. If -fstack-protector causes serious |
34 |
> problems and this is inherent to the nature of the package/etc then |
35 |
> just filter it. If it causes problems because upstream isn't doing |
36 |
> things right, then this is no different than how things were 10 years |
37 |
> ago when we were stomping out amd64 issues left and right (not working |
38 |
> on amd64 wasn't a reason to mask a package for x86, but we didn't |
39 |
> accept "upstream doesn't care about amd64" as an excuse). |
40 |
> |
41 |
> Staying close to upstream is a good philosophy in general. I don't |
42 |
> think that means that we can't have reasonable CFLAGS. Otherwise we |
43 |
> wouldn't set CFLAGS at all and would just use whatever defaults the |
44 |
> upstream build system applies. We're a distro - doing integration |
45 |
> work is a value-add, not a sin. If we aren't doing integration, then |
46 |
> users might as well just do Linux-from-scratch. |
47 |
> |
48 |
> Rich |
49 |
> |
50 |
|
51 |
Past events have led me to think that we are sometimes too dependent on |
52 |
upstream for guidance. I have certainly deviated from upstream whenever |
53 |
it made sense and the results have been fairly positive. I am not saying |
54 |
that acting for the best interests of Gentoo is mutually exclusive with |
55 |
collaboration with upstream, but I am saying that the two sometimes |
56 |
conflict. This being one of them. I am taking this opportunity to point |
57 |
out that what is best for Gentoo should come first and that it is okay |
58 |
to make decisions on our own. |