1 |
Peter Humphrey <prh@××××××××××.uk> posted |
2 |
200611051210.25023.prh@××××××××××.uk, excerpted below, on Sun, 05 Nov |
3 |
2006 12:10:24 +0000: |
4 |
|
5 |
[rewrapped] |
6 |
|
7 |
> $ emerge --info | grep FLAGS |
8 |
|
9 |
> CFLAGS="-march=k8 -Os -pipe -frename-registers -fweb -freorder-blocks |
10 |
> -freorder-blocks-and-partition -funit-at-a-time -fgcse-sm -fgcse-las |
11 |
> -fgcse-after-reload -fmerge-all-constants" |
12 |
|
13 |
> CXXFLAGS="-march=k8 -Os -pipe -frename-registers -fweb -freorder-blocks |
14 |
> -funit-at-a-time -fgcse-sm -fgcse-las -fgcse-after-reload |
15 |
> -fmerge-all-constants" |
16 |
|
17 |
> Unset: CTARGET, INSTALL_MASK, LDFLAGS, PORTAGE_RSYNC_EXTRA_OPTS |
18 |
|
19 |
> So I have a pointer away from glibc and linux-headers. CFLAGS on the new |
20 |
> system are as on the old but with -combine added.The old and new systems |
21 |
> have the same versions of kernel, linux-headers and glibc, but different |
22 |
> versions of gcc. Rsync is crippled on the old system but fine on the new |
23 |
> one. |
24 |
|
25 |
Well, I /did/ say gcc 4.1.1 seemed way (as in, vastly) better than 3.4.x, |
26 |
for me, so at least that's being borne out. |
27 |
|
28 |
As for CFLAGS, the only change I've made recently is adding |
29 |
-ftree-vectorize, experimentally, given the discussion here. I'm not sure |
30 |
it's the problem, particularly as where I /am/ experiencing issues the |
31 |
first thing I tried was compiling stuff without it, so if it's the |
32 |
problem, it's in some obscure dependency somewhere, but I certainly had no |
33 |
issues before trying it and now I do, so for anyone else thinking of |
34 |
trying it, I'd recommend staying well away from -ftree-vectorize for the |
35 |
time being. I'd really like to be able to say for sure it is the problem, |
36 |
and find it very frustrating not to be able to nail it down, but suffice |
37 |
it to say /something's/ causing me problems right now and that's one of my |
38 |
recent changes, so I'd recommend staying well away from it. |
39 |
|
40 |
I mentioned that I need to do an emerge -emptytree again, as it has been |
41 |
awhile. I'm still debating whether to try it with or without |
42 |
-ftree-vectorize. If that's the problem, it may well go away if |
43 |
everything is compiled with it. OTOH, it may not. That's an awful lot of |
44 |
compiling to do and still have a problem if -ftree-vectorize is causing it |
45 |
and it doesn't go away with everything compiled with it, but OTOH, if |
46 |
that's /not/ the problem, and I decide not to try it, I'll never know |
47 |
whether that was the problem or not. |
48 |
|
49 |
I'm about to decide to simply play it safe and pretend nobody here ever |
50 |
called -ftree-vectorize to my attention, at least until the flag has had a |
51 |
bit more time to mature (say gcc 4.2 or 4.3). Only just because I'm the |
52 |
type of person I am, that would bother me as I will have never nailed it |
53 |
for sure. In any case, my best guess right now are that the issues I am |
54 |
having are related to a nasty interplay between -ftree-vectorize and |
55 |
glibc-2.5. In looking around at Gentoo glibc bugs, I found at least one |
56 |
equally strange one, on x86. The bug required glibc 2.5, and it required |
57 |
that a certain package be compiled with a certain cflag. With that cflag |
58 |
on glibc 2.4, it worked fine, as it did without the cflag on 2.5, but put |
59 |
them together and things blew up. It wasn't -ftree-vectorize, and it was |
60 |
on x86, but I've a rather strong suspicion my case is similar, only on |
61 |
amd64 and with a different cflag, -ftree-vectorize most likely. I've just |
62 |
not figured out which particular package is doing it, and not being able |
63 |
to figure it out as I said is VERY annoying to me, WAY more so than the |
64 |
bit of inconvenience the actual bug is causing. (Yes, I know I said all |
65 |
that already, but it's still annoying and I'm still griping about it! =8^( |
66 |
|
67 |
-- |
68 |
Duncan - List replies preferred. No HTML msgs. |
69 |
"Every nonfree program has a lord, a master -- |
70 |
and if you use the program, he is your master." Richard Stallman |
71 |
|
72 |
-- |
73 |
gentoo-amd64@g.o mailing list |