| To: | Atsushi Nemoto <anemo@mba.ocn.ne.jp> |
|---|---|
| Subject: | Re: jiffies_64 vs. jiffies |
| From: | Paul Mackerras <paulus@samba.org> |
| Date: | Fri, 3 Mar 2006 15:06:16 +1100 |
| Cc: | linux-kernel@vger.kernel.org, linux-mips@linux-mips.org, akpm@osdl.org |
| In-reply-to: | <20060301.144442.118975101.nemoto@toshiba-tops.co.jp> |
| Original-recipient: | rfc822;linux-mips@linux-mips.org |
| References: | <20060301.144442.118975101.nemoto@toshiba-tops.co.jp> |
| Sender: | linux-mips-bounce@linux-mips.org |
Atsushi Nemoto writes: > Hi. I noticed that the 'jiffies' variable has 'wall_jiffies + 1' > value in most of time. I'm using MIPS platform but I think this is > same for other platforms. > > I suppose this is due to gcc does not know that jiffies_64 and jiffies > share same place. I can confirm that the same thing happens on powerpc, both 32-bit and 64-bit. The compiler loads up jiffies, jiffies_64 and wall_jiffies into registers before storing back the incremented value into jiffies_64 and then updating wall_jiffies. Thanks for finding that, it explains some other strange things that I have seen happen. Paul. |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | nfs network timeout on bcm1480 BigSur, dan.mcgee@ntsoc.com |
|---|---|
| Next by Date: | [PATCH] serial8250: set divisor register correctly for AMD Alchemy SoC uart, Jon Anders Haugum |
| Previous by Thread: | Re: jiffies_64 vs. jiffies, Atsushi Nemoto |
| Next by Thread: | [PATCH] [MIPS] use USECS_PER_SEC / HZ instead of tick_usec in do_gettimeofday, Atsushi Nemoto |
| Indexes: | [Date] [Thread] [Top] [All Lists] |