linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: Josh Boyer <jwboyer@linux.vnet.ibm.com>
To: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: linuxppc-dev@ozlabs.org
Subject: Re: [PATCH 8/25] powerpc: Improve support for 4xx indirect DCRs
Date: Thu, 6 Dec 2007 20:19:13 -0600	[thread overview]
Message-ID: <20071206201913.498185e6@zod.rchland.ibm.com> (raw)
In-Reply-To: <20071206080117.754FEDDE45@ozlabs.org>

On Thu, 06 Dec 2007 19:00:06 +1100
Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:

> Accessing indirect DCRs is done via a pair of address/data DCRs.
> 
> Such accesses are thus inherently racy, vs. interrupts, preemption
> and possibly SMP if 4xx SMP cores are ever used.
> 
> This updates the mfdcri/mtdcri macros in dcr-native.h (which were
> so far unused) to use a spinlock.
> 
> In addition, add some common definitions to a new dcr-regs.h file.
> 
> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> ---
> 
> 
>  arch/powerpc/sysdev/dcr.c        |    1 
>  include/asm-powerpc/dcr-native.h |   30 +++++++++++-----
>  include/asm-powerpc/dcr-regs.h   |   72 +++++++++++++++++++++++++++++++++++++++
>  include/asm-powerpc/dcr.h        |    1 
>  4 files changed, 96 insertions(+), 8 deletions(-)
> 
> Index: linux-work/arch/powerpc/sysdev/dcr.c
> ===================================================================
> --- linux-work.orig/arch/powerpc/sysdev/dcr.c	2007-11-28 13:34:45.000000000 +1100
> +++ linux-work/arch/powerpc/sysdev/dcr.c	2007-11-28 13:34:48.000000000 +1100
> @@ -139,3 +139,4 @@ void dcr_unmap(dcr_host_t host, unsigned
>  EXPORT_SYMBOL_GPL(dcr_unmap);
> 
>  #endif /* !defined(CONFIG_PPC_DCR_NATIVE) */
> +

Somehow I doubt this single extraneous whitespace addition is really
needed.


> Index: linux-work/include/asm-powerpc/dcr-native.h
> ===================================================================
> --- linux-work.orig/include/asm-powerpc/dcr-native.h	2007-11-28 13:33:51.000000000 +1100
> +++ linux-work/include/asm-powerpc/dcr-native.h	2007-11-28 15:22:22.000000000 +1100
> @@ -22,6 +22,8 @@
>  #ifdef __KERNEL__
>  #ifndef __ASSEMBLY__
> 
> +#include <linux/spinlock.h>
> +
>  typedef struct {
>  	unsigned int base;
>  } dcr_host_t;
> @@ -55,18 +57,30 @@ do {								\
>  } while (0)
> 
>  /* R/W of indirect DCRs make use of standard naming conventions for DCRs */
> -#define mfdcri(base, reg)			\
> -({						\
> -	mtdcr(base ## _CFGADDR, base ## _ ## reg);	\
> -	mfdcr(base ## _CFGDATA);			\
> +extern spinlock_t dcr_ind_lock;
> +
> +#define mfdcri(base, reg)				\
> +({							\
> +	unsigned long flags; 				\
> +	unsigned int val;				\
> +	spin_lock_irqsave(&dcr_ind_lock, flags);	\
> +	mtdcr(DCRN_ ## base ## _CONFIG_ADDR, reg);	\
> +	val = mfdcr(DCRN_ ## base ## _CONFIG_DATA);	\
> +	spin_unlock_irqrestore(&dcr_ind_lock, flags);	\
> +	val;						\
>  })
> 
> -#define mtdcri(base, reg, data)			\
> -do {						\
> -	mtdcr(base ## _CFGADDR, base ## _ ## reg);	\
> -	mtdcr(base ## _CFGDATA, data);		\
> +#define mtdcri(base, reg, data)				\
> +do {							\
> +	unsigned long flags; 				\
> +	spin_lock_irqsave(&dcr_ind_lock, flags);	\
> +	mtdcr(DCRN_ ## base ## _CONFIG_ADDR, reg);	\
> +	mtdcr(DCRN_ ## base ## _CONFIG_DATA, data);	\
> +	spin_unlock_irqrestore(&dcr_ind_lock, flags);	\
>  } while (0)
> 
> +
> +

More bogus whitespace.


>  #endif /* __ASSEMBLY__ */
>  #endif /* __KERNEL__ */
>  #endif /* _ASM_POWERPC_DCR_NATIVE_H */
> Index: linux-work/include/asm-powerpc/dcr.h
> ===================================================================
> --- linux-work.orig/include/asm-powerpc/dcr.h	2007-11-28 13:40:13.000000000 +1100
> +++ linux-work/include/asm-powerpc/dcr.h	2007-11-28 13:49:37.000000000 +1100
> @@ -40,6 +40,7 @@ extern unsigned int dcr_resource_len(str
>  				     unsigned int index);
>  #endif /* CONFIG_PPC_MERGE */
> 
> +

You really like whitespace.  Again I don't see anything needed for this
file :)

>  #endif /* CONFIG_PPC_DCR */
>  #endif /* __KERNEL__ */
>  #endif /* _ASM_POWERPC_DCR_H */
> Index: linux-work/include/asm-powerpc/dcr-regs.h
> ===================================================================
> --- /dev/null	1970-01-01 00:00:00.000000000 +0000
> +++ linux-work/include/asm-powerpc/dcr-regs.h	2007-11-28 14:43:25.000000000 +1100
> @@ -0,0 +1,72 @@
> +/*
> + * Common DCR / SDR / CPR register definitions used on various IBM/AMCC
> + * 4xx processors
> + *
> + *    Copyright 2007 Benjamin Herrenschmidt, IBM Corp
> + *                   <benh@kernel.crashing.org>
> + *
> + * Mostly lifted from asm-ppc/ibm4xx.h by
> + *
> + *    Copyright (c) 1999 Grant Erickson <grant@lcse.umn.edu>
> + *
> + */
> +
> +#ifndef __DCR_REGS_H__
> +#define __DCR_REGS_H__
> +
> +/*
> + * Most DCRs used for controlling devices such as the MAL, DMA engine,
> + * etc... are obtained for the device tree.
> + *
> + * The definitions in this files are fixed DCRs and indirect DCRs that
> + * are commonly used outside of specific drivers or refer to core
> + * common registers that may occasionally have to be tweaked outside
> + * of the driver main register set
> + */
> +
> +/* CPRs (440GX and 440SP/440SPe) */
> +#define DCRN_CPR0_CONFIG_ADDR	0xc
> +#define DCRN_CPR0_CONFIG_DATA	0xd
> +
> +/* SDRs (440GX and 440SP/440SPe) */
> +#define DCRN_SDR0_CONFIG_ADDR 	0xe
> +#define DCRN_SDR0_CONFIG_DATA	0xf
> +
> +#define SDR0_PFC0		0x4100
> +#define SDR0_PFC1		0x4101
> +#define SDR0_PFC1_EPS		0x1c00000
> +#define SDR0_PFC1_EPS_SHIFT	22
> +#define SDR0_PFC1_RMII		0x02000000
> +#define SDR0_MFR		0x4300
> +#define SDR0_MFR_TAH0 		0x80000000  	/* TAHOE0 Enable */
> +#define SDR0_MFR_TAH1 		0x40000000  	/* TAHOE1 Enable */
> +#define SDR0_MFR_PCM  		0x10000000  	/* PPC440GP irq compat mode */
> +#define SDR0_MFR_ECS  		0x08000000  	/* EMAC int clk */
> +#define SDR0_MFR_T0TXFL		0x00080000
> +#define SDR0_MFR_T0TXFH		0x00040000
> +#define SDR0_MFR_T1TXFL		0x00020000
> +#define SDR0_MFR_T1TXFH		0x00010000
> +#define SDR0_MFR_E0TXFL		0x00008000
> +#define SDR0_MFR_E0TXFH		0x00004000
> +#define SDR0_MFR_E0RXFL		0x00002000
> +#define SDR0_MFR_E0RXFH		0x00001000
> +#define SDR0_MFR_E1TXFL		0x00000800
> +#define SDR0_MFR_E1TXFH		0x00000400
> +#define SDR0_MFR_E1RXFL		0x00000200
> +#define SDR0_MFR_E1RXFH		0x00000100
> +#define SDR0_MFR_E2TXFL		0x00000080
> +#define SDR0_MFR_E2TXFH		0x00000040
> +#define SDR0_MFR_E2RXFL		0x00000020
> +#define SDR0_MFR_E2RXFH		0x00000010
> +#define SDR0_MFR_E3TXFL		0x00000008
> +#define SDR0_MFR_E3TXFH		0x00000004
> +#define SDR0_MFR_E3RXFL		0x00000002
> +#define SDR0_MFR_E3RXFH		0x00000001
> +#define SDR0_UART0		0x0120
> +#define SDR0_UART1		0x0121
> +#define SDR0_UART2		0x0122
> +#define SDR0_UART3		0x0123
> +#define SDR0_CUST0		0x4000
> +
> +
> +#endif /* __DCR_REGS_H__ */

  reply	other threads:[~2007-12-07  2:22 UTC|newest]

Thread overview: 60+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-12-06  7:59 [PATCH 0/25] powerpc: 4xx PCI, PCI-X and PCI-Express support among others Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 2/25] powerpc: Merge pci_process_bridge_OF_ranges() Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 1/25] powerpc: Make isa_mem_base common to 32 and 64 bits Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 3/25] powerpc: Fix powerpc 32 bits resource fixup for 64 bits resources Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 4/25] powerpc: Reworking machine check handling and Fix 440/440A Benjamin Herrenschmidt
2007-12-10 17:59   ` Josh Boyer
2007-12-10 20:33     ` Benjamin Herrenschmidt
2007-12-10 20:44       ` Josh Boyer
2007-12-06  8:00 ` [PATCH 5/25] powerpc: Add xmon function to dump 44x TLB Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 6/25] powerpc: Change 32 bits PCI message about resource allocation Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 7/25] powerpc: Add of_translate_dma_address Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 8/25] powerpc: Improve support for 4xx indirect DCRs Benjamin Herrenschmidt
2007-12-07  2:19   ` Josh Boyer [this message]
2007-12-10  4:54     ` Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 9/25] powerpc: 4xx PLB to PCI-X support Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 10/25] powerpc: 4xx PLB to PCI 2.x support Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 11/25] powerpc: 4xx PLB to PCI Express support Benjamin Herrenschmidt
2007-12-06  9:26   ` Stefan Roese
2007-12-06 22:21     ` Benjamin Herrenschmidt
2007-12-07  7:02       ` Stefan Roese
2007-12-09  7:07         ` Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 12/25] powerpc: PCI support for 4xx Ebony board Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 13/25] powerpc: Remove useless volatiles in udbg_16550.c Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 14/25] powerpc: Add early udbg support for 40x processors Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 15/25] powerpc: early debug forces console log level to max Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 16/25] powerpc: EP405 boards support for arch/powerpc Benjamin Herrenschmidt
2007-12-07  3:05   ` Josh Boyer
2007-12-09  6:57     ` Benjamin Herrenschmidt
2007-12-10  4:56     ` Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 17/25] powerpc: Add PCI to Walnut platform Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 18/25] powerpc: Base support for 440GX Taishan eval board Benjamin Herrenschmidt
2007-12-06 19:56   ` Josh Boyer
2007-12-07  3:17   ` Josh Boyer
2007-12-09  7:01     ` Benjamin Herrenschmidt
2007-12-09 17:24       ` Olof Johansson
2007-12-09 18:24         ` Josh Boyer
2007-12-09 20:20           ` Olof Johansson
2007-12-09 19:58         ` Benjamin Herrenschmidt
2007-12-10  1:29           ` Josh Boyer
2007-12-09 23:43         ` CC munging by mailman lists (Was: Re: [PATCH 18/25] powerpc: Base support for 440GX Taishan eval board) Stephen Rothwell
2007-12-10  5:39     ` [PATCH 18/25] powerpc: Base support for 440GX Taishan eval board Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 20/25] powerpc: Wire up 440EP USB controlle support to Bamboo board Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 19/25] powerpc: Wire up PCI on " Benjamin Herrenschmidt
2007-12-07  3:19   ` Josh Boyer
2007-12-09  7:03     ` Benjamin Herrenschmidt
2007-12-10  5:40     ` Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 21/25] powerpc: Adds decoding of 440SPE memory size to boot wrapper library Benjamin Herrenschmidt
2007-12-07  3:22   ` Josh Boyer
2007-12-09  7:04     ` Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 22/25] powerpc: Add mfspr/mtspr inline macros to 4xx bootwrapper Benjamin Herrenschmidt
2007-12-06  8:00 ` [PATCH 23/25] powerpc: Rework 4xx clock probing in boot wrapper Benjamin Herrenschmidt
2007-12-07  3:27   ` Josh Boyer
2007-12-09  7:05     ` Benjamin Herrenschmidt
2007-12-10  5:43     ` Benjamin Herrenschmidt
2007-12-10 11:54       ` Josh Boyer
2007-12-06  8:00 ` [PATCH 24/25] powerpc: Base support for 440SPe "Katmai" eval board Benjamin Herrenschmidt
2007-12-06  9:42   ` Stefan Roese
2007-12-06  8:00 ` [PATCH 25/25] powerpc: 4xx PCI-E Link setup improvements Benjamin Herrenschmidt
2007-12-06 13:32 ` [PATCH 0/25] powerpc: 4xx PCI, PCI-X and PCI-Express support among others Josh Boyer
2007-12-06 20:25 ` Josh Boyer

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20071206201913.498185e6@zod.rchland.ibm.com \
    --to=jwboyer@linux.vnet.ibm.com \
    --cc=benh@kernel.crashing.org \
    --cc=linuxppc-dev@ozlabs.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).