From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: Valentine Barshak <vbarshak@ru.mvista.com>
Cc: linuxppc-dev@ozlabs.org
Subject: Re: [RFC][PATCH] PowerPC: 4xx PCIe indirect DCR spinlock fix.
Date: Tue, 05 Feb 2008 07:50:39 +1100 [thread overview]
Message-ID: <1202158239.7079.4.camel@pasglop> (raw)
In-Reply-To: <20080204182740.GA2898@ru.mvista.com>
On Mon, 2008-02-04 at 21:27 +0300, Valentine Barshak wrote:
> Since we have mfdcri() and mtdcri() as macros, we can't use constructions, such
> as "mtdcri(base, reg, mfdcri(base, reg) | val)". In this case the mfdcri() stuff
> is not evaluated first. It's evaluated inside the mtdcr() macro and we have
> the dcr_ind_lock spinlock acquired twice. To avoid this error, I've added
> set_dcri() and clr_dcri() macros which set/clear the specified bits.
>
> Signed-off-by: Valentine Barshak <vbarshak@ru.mvista.com>
No, better is to fix the macros to use functions.
Make mfdcri/mtdcri call into __mfdcri/__mtdcri functions after fixing
up the macro register names, and have those be inlines that take the
lock.
Ben.
> arch/powerpc/sysdev/ppc4xx_pci.c | 13 +++++--------
> include/asm-powerpc/dcr-native.h | 22 ++++++++++++++++++++++
> 2 files changed, 27 insertions(+), 8 deletions(-)
>
> diff -pruN linux-2.6.orig/arch/powerpc/sysdev/ppc4xx_pci.c linux-2.6/arch/powerpc/sysdev/ppc4xx_pci.c
> --- linux-2.6.orig/arch/powerpc/sysdev/ppc4xx_pci.c 2008-02-04 20:05:37.000000000 +0300
> +++ linux-2.6/arch/powerpc/sysdev/ppc4xx_pci.c 2008-02-04 20:16:33.000000000 +0300
> @@ -645,7 +645,7 @@ static int __init ppc440spe_pciex_core_i
> int time_out = 20;
>
> /* Set PLL clock receiver to LVPECL */
> - mtdcri(SDR0, PESDR0_PLLLCT1, mfdcri(SDR0, PESDR0_PLLLCT1) | 1 << 28);
> + set_dcri(SDR0, PESDR0_PLLLCT1, 1 << 28);
>
> /* Shouldn't we do all the calibration stuff etc... here ? */
> if (ppc440spe_pciex_check_reset(np))
> @@ -659,8 +659,7 @@ static int __init ppc440spe_pciex_core_i
> }
>
> /* De-assert reset of PCIe PLL, wait for lock */
> - mtdcri(SDR0, PESDR0_PLLLCT1,
> - mfdcri(SDR0, PESDR0_PLLLCT1) & ~(1 << 24));
> + clr_dcri(SDR0, PESDR0_PLLLCT1, 1 << 24);
> udelay(3);
>
> while (time_out) {
> @@ -712,9 +711,8 @@ static int ppc440spe_pciex_init_port_hw(
> mtdcri(SDR0, port->sdr_base + PESDRn_440SPE_HSSL7SET1,
> 0x35000000);
> }
> - val = mfdcri(SDR0, port->sdr_base + PESDRn_RCSSET);
> - mtdcri(SDR0, port->sdr_base + PESDRn_RCSSET,
> - (val & ~(1 << 24 | 1 << 16)) | 1 << 12);
> + clr_dcri(SDR0, port->sdr_base + PESDRn_RCSSET, 1 << 24 | 1 << 16)
> + set_dcri(SDR0, port->sdr_base + PESDRn_RCSSET, 1 << 12);
>
> return 0;
> }
> @@ -1042,8 +1040,7 @@ static int __init ppc4xx_pciex_port_init
> port->link = 0;
> }
>
> - mtdcri(SDR0, port->sdr_base + PESDRn_RCSSET,
> - mfdcri(SDR0, port->sdr_base + PESDRn_RCSSET) | 1 << 20);
> + set_dcri(SDR0, port->sdr_base + PESDRn_RCSSET, 1 << 20);
> msleep(100);
>
> return 0;
> diff -pruN linux-2.6.orig/include/asm-powerpc/dcr-native.h linux-2.6/include/asm-powerpc/dcr-native.h
> --- linux-2.6.orig/include/asm-powerpc/dcr-native.h 2008-02-04 20:06:34.000000000 +0300
> +++ linux-2.6/include/asm-powerpc/dcr-native.h 2008-02-04 20:16:33.000000000 +0300
> @@ -79,6 +79,28 @@ do { \
> spin_unlock_irqrestore(&dcr_ind_lock, flags); \
> } while (0)
>
> +#define set_dcri(base, reg, data) \
> +do { \
> + 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); \
> + mtdcr(DCRN_ ## base ## _CONFIG_DATA, val | (data)); \
> + spin_unlock_irqrestore(&dcr_ind_lock, flags); \
> +} while (0)
> +
> +#define clr_dcri(base, reg, data) \
> +do { \
> + 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); \
> + mtdcr(DCRN_ ## base ## _CONFIG_DATA, val & ~(data)); \
> + spin_unlock_irqrestore(&dcr_ind_lock, flags); \
> +} while (0)
> +
> #endif /* __ASSEMBLY__ */
> #endif /* __KERNEL__ */
> #endif /* _ASM_POWERPC_DCR_NATIVE_H */
next prev parent reply other threads:[~2008-02-04 20:50 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-02-04 18:27 [RFC][PATCH] PowerPC: 4xx PCIe indirect DCR spinlock fix Valentine Barshak
2008-02-04 20:50 ` Benjamin Herrenschmidt [this message]
2008-02-05 18:36 ` Valentine Barshak
2008-02-05 20:47 ` Benjamin Herrenschmidt
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=1202158239.7079.4.camel@pasglop \
--to=benh@kernel.crashing.org \
--cc=linuxppc-dev@ozlabs.org \
--cc=vbarshak@ru.mvista.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.