Messages in this thread | | | Date | Sat, 28 Nov 2009 00:41:27 +0100 | From | Jiri Slaby <> | Subject | [PATCH v3 00/27] writable limits |
| |
Hi,
I broke the threading to not mess up with the long thread.
In this version I got rid of the rlim access_only ugliness.
There are two things: 1) <quote from=Ingo> A prominent example would be the stack limit - we base address layout decisions on it. Check arch/x86/mm/mmap.c. RLIM_INFINITY has a special meaning plus we also set mmap_base() based on the rlim. </quote> Should there be some special handling of that? In standard setrlimit there is none.
2) <quote from=Oleg> Hmm. you are right. Do you know why acct_file_reopen() does if (old_acct) do_acct_process(); ??? </quote> As I expressed myself before, I don't know why it is there (it doesn't make sense to me either). But I took a look at when it was added. From the very first merge of acct.c (2.1.68pre1) it was just in (name == NULL) path (turning acct off). Then in 2.1.126 it was switched to account on every accounting file change.
I fear if we changed this, something would break.
thanks, -- js
| |