Re: [Uclinux-dist-devel] [PATCH 1/2] mtd: m25p80: Reworkprobing/JEDEC code
From: Barry Song <hidden>
Date: 2010-06-21 03:27:54
Also in:
lkml
On Mon, Jun 21, 2010 at 10:42 AM, Song, Barry [off-list ref] wrote= :
quoted
-----Original Message----- From: uclinux-dist-devel-bounces@blackfin.uclinux.org [mailto:uclinux-dist-devel-bounces@blackfin.uclinux.org] On Behalf Of Anton Vorontsov Sent: Friday, June 18, 2010 9:32 PM To: Barry Song Cc: David Brownell; Artem Bityutskiy; linux-kernel@vger.kernel.org; linuxppc-dev@ozlabs.org; linux-mtd@lists.infradead.org; uclinux-dist-devel@blackfin.uclinux.org; Andrew Morton Subject: Re: [Uclinux-dist-devel] [PATCH 1/2] mtd: m25p80: Reworkprobing/JEDEC code On Sat, Jun 12, 2010 at 02:27:12PM +0800, Barry Song wrote:quoted
On Wed, Aug 19, 2009 at 5:46 AM, Anton Vorontsov [off-list ref] wrote:quoted
Previosly the driver always tried JEDEC probing, assumingthat non-JEDECquoted
quoted
chips will return '0'. But truly non-JEDEC chips (likeCAT25) won't doquoted
quoted
that, their behaviour on RDID command is undefined, so thedriver shouldquoted
quoted
not call jedec_probe() for these chips. Also, be less strict on error conditions, don't fail toprobe if JEDECquoted
quoted
found a chip that is different from what platform codetold, insteadquoted
quoted
just print some warnings and use an information obtainedvia JEDEC. Inquoted
This patch caused a problem: even though the external flash doesn't exist, it will still pass the probe() and be registerred into kernel and given the partition table. You may refer to this bug report:http://blackfin.uclinux.org/gf/project/uclinux-dist/tracker/?ac tion=3DTrackerItemEdit&tracker_item_id=3D5975&start=3D0 Thanks for the report. There's little we can do about it. Platform code asked us to register the device, and JEDEC probing of M25Pxx chips isn't reliable (thanks to various vendors that make these JEDEC and non-JEDEC variants), so the best thing we can do is to register the chip anyway. OTOH, if the board pulls MISO line up, then the following patch should help.Make sense with pullup to keep the value high while external device doesn't exist.quoted
If this won't work, we'll have to add some flag to the platform data, i.e. to force JEDEC probing, and not trust platform data.How about we add a non_jedec flag in platform_data, if the flag is 1, we let the detection pass even though the ID is 0? Otherwise, we need a valid ID?
Here i mean: Index: drivers/mtd/devices/m25p80.c =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
--- drivers/mtd/devices/m25p80.c (revision 8927)
+++ drivers/mtd/devices/m25p80.c (revision 8929)@@ -795,8 +795,13 @@ jid =3D jedec_probe(spi); if (!jid) { - dev_info(&spi->dev, "non-JEDEC variant of %s\n", - id->name); + if (!data->non_jedec) { + dev_err(&spi->dev, "fail to detect%s\n", + id->name); + return -ENODEV; + } else + dev_info(&spi->dev, "non-JEDEC variant of %s\n", + id->name); } else if (jid !=3D id) { /* * JEDEC knows better, so overwrite platform ID. We
Index: include/linux/spi/flash.h =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
--- include/linux/spi/flash.h (revision 8927)
+++ include/linux/spi/flash.h (revision 8929)@@ -25,6 +25,11 @@ char *type; + /* + * For non-JEDEC, id will be 0. In this case, we can't be sure + * whether the flash exists with runtime probing. + */ + int non_jedec; /* we'll likely add more ... use JEDEC IDs, etc */ };
quoted
Not-yet-Signed-off-by: Anton Vorontsov [off-list ref] ---diff --git a/drivers/mtd/devices/m25p80.cb/drivers/mtd/devices/m25p80.c index 81e49a9..a307929 100644--- a/drivers/mtd/devices/m25p80.c +++ b/drivers/mtd/devices/m25p80.c@@ -16,6 +16,7 @@=C2=A0*/ #include <linux/init.h> +#include <linux/errno.h> #include <linux/module.h> #include <linux/device.h> #include <linux/interrupt.h>@@ -723,7 +724,7 @@ static const struct spi_device_id*__devinit jedec_probe(struct spi_device *spi) =C2=A0 =C2=A0 =C2=A0 if (tmp < 0) { =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 DEBUG(MTD_DEBUG_LEVEL0,=
"%s: error %d reading
quoted
JEDEC ID\n", =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 dev_name(&spi->dev), tmp);
quoted
- =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0return NULL; + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0return ERR_PTR(tmp); =C2=A0 =C2=A0 =C2=A0 } =C2=A0 =C2=A0 =C2=A0 jedec =3D id[0]; =C2=A0 =C2=A0 =C2=A0 jedec =3D jedec << 8;@@ -737,7 +738,7 @@ static const struct spi_device_id*__devinit jedec_probe(struct spi_device *spi) =C2=A0 =C2=A0 =C2=A0 =C2=A0* exist for non-JEDEC chips, but for compatib=
ility
quoted
they return ID 0. =C2=A0 =C2=A0 =C2=A0 =C2=A0*/ =C2=A0 =C2=A0 =C2=A0 if (jedec =3D=3D 0) - =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0return NULL; + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0return ERR_PTR(-EEXIST)=
;
quoted
=C2=A0 =C2=A0 =C2=A0 ext_jedec =3D id[3] << 8 | id[4];@@ -749,7 +750,7 @@ static const struct spi_device_id*__devinit jedec_probe(struct spi_device *spi) =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 return &m25p_ids[tmp];
quoted
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 } =C2=A0 =C2=A0 =C2=A0 } - =C2=A0 =C2=A0 =C2=A0return NULL; + =C2=A0 =C2=A0 =C2=A0return ERR_PTR(-ENODEV); }@@ -794,9 +795,11 @@ static int __devinit m25p_probe(structspi_device *spi) =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 const struct spi_device=
_id *jid;
quoted
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 jid =3D jedec_probe(spi=
);
quoted
- =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0if (!jid) { + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0if (IS_ERR(jid) && PTR_=
ERR(jid) =3D=3D -EEXIST) {quoted
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 dev_info(&spi->dev, "non-JEDEC variant of %s\n",
quoted
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0id->name);
quoted
+ =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0} else if (IS_ERR(jid))=
{quoted
+ =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0return PTR_ERR(jid);
quoted
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 } else if (jid !=3D id)=
{quoted
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 /*
quoted
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 =C2=A0* JEDEC knows better, so overwrite
quoted
platform ID. We _______________________________________________ Uclinux-dist-devel mailing list Uclinux-dist-devel@blackfin.uclinux.org https://blackfin.uclinux.org/mailman/listinfo/uclinux-dist-devel