| To: | Pete Popov <ppopov@embeddedalley.com> |
|---|---|
| Subject: | Re: mdelay() from board_setup() [is default value for loops_per_jiffy way off?] |
| From: | Ralf Baechle <ralf@linux-mips.org> |
| Date: | Wed, 20 Apr 2005 19:07:20 +0100 |
| Cc: | Clem Taylor <clem.taylor@gmail.com>, linux-mips@linux-mips.org |
| In-reply-to: | <42669862.7000405@embeddedalley.com> |
| Original-recipient: | rfc822;linux-mips@linux-mips.org |
| References: | <ecb4efd10504201050a00f941@mail.gmail.com> <42669862.7000405@embeddedalley.com> |
| Sender: | linux-mips-bounce@linux-mips.org |
| User-agent: | Mutt/1.4.1i |
On Wed, Apr 20, 2005 at 10:58:58AM -0700, Pete Popov wrote: > It's too early in board_setup() to use the standard delay routines. You > can't use those until after calibrate_delay() runs. To do precise delays in > board_setup, you'll have to do something yourself where you read the cp0 > timer periodically and wait a certain amount of time. And make sure not to be trapped by wrap-arounds of that counter. Also it doesn't count the same speed on all processors ... Ralf |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: mdelay() from board_setup() [is default value for loops_per_jiffy way off?], Pete Popov |
|---|---|
| Next by Date: | Re: Bug detection and correction on Alchemy au1x00_uart.c serial driver, Pete Popov |
| Previous by Thread: | Re: mdelay() from board_setup() [is default value for loops_per_jiffy way off?], Pete Popov |
| Next by Thread: | troubles writing to a mmapped PCI BAR, Clem Taylor |
| Indexes: | [Date] [Thread] [Top] [All Lists] |