| 1 |
Hi, |
| 2 |
|
| 3 |
On Sat, 13 Jan 2007 12:29:09 -0500 |
| 4 |
Randy Barlow <randy@×××××××××××××××××.com> wrote: |
| 5 |
|
| 6 |
> On Saturday 13 January 2007 09:42, Michael Sullivan wrote: |
| 7 |
> > This strace doesn't help me much. What does "attached" mean, anyway? |
| 8 |
> |
| 9 |
> I don't know what the problem is, but I can tell you that strace "attaches" to |
| 10 |
> a process meaning that it begins to watch the process to see what system |
| 11 |
> calls it is making. If the process doesn't make any system calls, then you |
| 12 |
> won't see anything with strace. [...] |
| 13 |
|
| 14 |
I think since the suggestion of using "strace" to look if a certain |
| 15 |
process is "doing" things was mine. So I owe this thread a bit more |
| 16 |
input. Randy, you're completely right. And especially the compile task |
| 17 |
in question, "gcc -o insn-attrtab.o" (shortened), is rather CPU- but |
| 18 |
not kernel-intensive. I don't know what exactly it does (not familiar |
| 19 |
with gcc internals), but it _heavily_ depends on the optimization |
| 20 |
level. So my suggestion to the OP would be to carefully look at the |
| 21 |
CFLAGS. Maybe -- please correct me if that is the case -- is a gcc |
| 22 |
build restricted to certain CFLAGS, though. At this stage, I would |
| 23 |
exclude a race condition involving the kernel, e.g. some stale or |
| 24 |
missing files. It may have to do with threading, but gcc doesn't thread. |
| 25 |
|
| 26 |
Question to the OP: What is the larger context of that gcc build? A |
| 27 |
simple update? A larger scale configuration change? |
| 28 |
|
| 29 |
-hwh |
| 30 |
-- |
| 31 |
gentoo-user@g.o mailing list |
Archives