| To: | sam@ravnborg.org |
|---|---|
| Subject: | Re: [PATCH] MIPS: Run checksyscalls for N32 and O32 ABI |
| From: | Atsushi Nemoto <anemo@mba.ocn.ne.jp> |
| Date: | Sun, 13 May 2007 22:51:54 +0900 (JST) |
| Cc: | vagabon.xyz@gmail.com, linux-mips@linux-mips.org, ralf@linux-mips.org |
| In-reply-to: | <20070512185854.GA8647@uranus.ravnborg.org> |
| Original-recipient: | rfc822;linux-mips@linux-mips.org |
| References: | <cda58cb80705111223y5e94eafcy710b878517c48c48@mail.gmail.com> <20070513.014713.74753298.anemo@mba.ocn.ne.jp> <20070512185854.GA8647@uranus.ravnborg.org> |
| Sender: | linux-mips-bounce@linux-mips.org |
On Sat, 12 May 2007 20:58:54 +0200, Sam Ravnborg <sam@ravnborg.org> wrote: > > Subject: [PATCH] MIPS: Simplify missing-syscalls for N32 and O32 > > This is overengineered. The only reason to make the syscall check > for each and every build was that this was easy and the missing syscalls > are easy to spot during a normal build. > But checking all combinations is just not worth it. > The arch responsible are assumed to build for the different architectures > once in a while so a missing syscall are likely to be detected anyway. > > We cannot run each and every consistency check in all combinations > for each build - that would end in only build noise. Well, 64-bit MIPS has three ABIs and each ABI has complete set of syscalls. So a result of default "missing-syscalls" target (which is for N64 ABI) is not useful at all for other two ABIs. I think checking them is worth even if the S/N ratio was quite low. --- Atsushi Nemoto |
| Previous by Date: | Re: [PATCH] MIPS: Run checksyscalls for N32 and O32 ABI, Sam Ravnborg |
|---|---|
| Next by Date: | Re: [PATCH] MIPS: Run checksyscalls for N32 and O32 ABI, Franck Bui-Huu |
| Previous by Thread: | Re: [PATCH] MIPS: Run checksyscalls for N32 and O32 ABI, Sam Ravnborg |
| Next by Thread: | Re: [PATCH] MIPS: Run checksyscalls for N32 and O32 ABI, Franck Bui-Huu |
| Indexes: | [Date] [Thread] [Top] [All Lists] |