Message ID | 51E3CB64.4080107@wanadoo.fr |
---|---|
State | New, archived |
Headers | show |
+ Marek, since he's been reviewing (with dismay?) the increase in macro flags in this driver. If there are any objections, I can amend/drop the patch. But in the meantime, I fixed a conflict with stuff that's been applied since you submitted this patch, and applied to l2-mtd.git. Thanks! Brian On Mon, Jul 15, 2013 at 12:13:56PM +0200, Michel Stempin wrote: > Add support for PMC (now Chingis, part of ISSI) Pm25LV512 (512 kBbit), > Pm25LV010 (1 Mbit) and Pm25LQ032 (32 Mbit) SPI Flash chips. > > This patch addresses two generations of PMC SPI Flash chips: > > - Pm25LV512 and Pm25LV010: these have 4KB sectors and 32KB > blocks. The 4KB sector erase uses a non-standard opcode > (0xd7). They do not support JEDEC RDID (0x9f), and so they can only > be detected by matching their name string with pre-configured > platform data. Because of the cascaded acquisitions, the datasheet > is no longer available on the current manufacturer's website, > although it is still commonly used in some recent wireless routers > (<https://forum.openwrt.org/viewtopic.php?pid=186360#p186360>). The > only public datasheet available seems to be on GeoCities: > <http://www.geocities.jp/scottle556/pdf/Pm25LV512-010.pdf> > > - Pm25LQ032: a newer generation flash, with 4KB sectors and 32KB > blocks. It uses the standard erase and JEDEC read-ID > opcodes. Manufacturer's datasheet is here: > <http://www.chingistek.com/img/Product_Files/Pm25LQ032C%20datasheet%20v1.6.1.pdf> > > This patch is resent in order to take into account both Brian Norris > remarks and this upstream patch: > > commit e534ee4f9ca29fdb38eea4b0c53f2154fbd8c1ee > Author: Krzysztof Mazur <krzysiek@podlesie.net> > Date: Fri Feb 22 15:51:05 2013 +0100 > > mtd: m25p80: introduce SST_WRITE flag for SST byte programming > > Not all SST devices implement the SST byte programming command. > Some devices (like SST25VF064C) implement only standard m25p80 page > write command. > > Now SPI flash devices that need sst_write() are explicitly marked > with new SST_WRITE flag and the decision to use sst_write() is based > on this flag instead of manufacturer id. > > Signed-off-by: Krzysztof Mazur <krzysiek@podlesie.net> > Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> > > Signed-off-by: Gabor Juhos <juhosg@openwrt.org> > Signed-off-by: Michel Stempin <michel.stempin@wanadoo.fr> > --- > drivers/mtd/devices/m25p80.c | 10 ++++++++++ > 1 file changed, 10 insertions(+) > > diff --git a/drivers/mtd/devices/m25p80.c b/drivers/mtd/devices/m25p80.c > index 2f3d2a5..55c8b5f 100644 > --- a/drivers/mtd/devices/m25p80.c > +++ b/drivers/mtd/devices/m25p80.c > @@ -43,6 +43,7 @@ > #define OPCODE_FAST_READ 0x0b /* Read data bytes (high frequency) */ > #define OPCODE_PP 0x02 /* Page program (up to 256 bytes) */ > #define OPCODE_BE_4K 0x20 /* Erase 4KiB block */ > +#define OPCODE_BE_4K_PMC 0xd7 /* Erase 4KiB block on PMC chips */ > #define OPCODE_BE_32K 0x52 /* Erase 32KiB block */ > #define OPCODE_CHIP_ERASE 0xc7 /* Erase whole flash chip */ > #define OPCODE_SE 0xd8 /* Sector erase (usually 64KiB) */ > @@ -682,6 +683,7 @@ struct flash_info { > #define SECT_4K 0x01 /* OPCODE_BE_4K works uniformly */ > #define M25P_NO_ERASE 0x02 /* No erase command needed */ > #define SST_WRITE 0x04 /* use SST byte programming */ > +#define SECT_4K_PMC 0x08 /* OPCODE_BE_4K_PMC works uniformly */ > }; > > #define INFO(_jedec_id, _ext_id, _sector_size, _n_sectors, _flags) \ > @@ -762,6 +764,11 @@ static const struct spi_device_id m25p_ids[] = { > { "n25q128a13", INFO(0x20ba18, 0, 64 * 1024, 256, 0) }, > { "n25q256a", INFO(0x20ba19, 0, 64 * 1024, 512, SECT_4K) }, > > + /* PMC */ > + { "pm25lv512", INFO(0, 0, 32 * 1024, 2, SECT_4K_PMC) }, > + { "pm25lv010", INFO(0, 0, 32 * 1024, 4, SECT_4K_PMC) }, > + { "pm25lq032", INFO(0x7f9d46, 0, 64 * 1024, 64, SECT_4K) }, > + > /* Spansion -- single (large) sector size only, at least > * for the chips listed here (without boot sectors). > */ > @@ -1014,6 +1021,9 @@ static int m25p_probe(struct spi_device *spi) > if (info->flags & SECT_4K) { > flash->erase_opcode = OPCODE_BE_4K; > flash->mtd.erasesize = 4096; > + } else if (info->flags & SECT_4K_PMC) { > + flash->erase_opcode = OPCODE_BE_4K_PMC; > + flash->mtd.erasesize = 4096; > } else { > flash->erase_opcode = OPCODE_SE; > flash->mtd.erasesize = info->sector_size;
Dear Brian Norris, > + Marek, since he's been reviewing (with dismay?) the increase in macro > flags in this driver. If there are any objections, I can amend/drop the > patch. Hmmm ... this SECT_4K_PMC seems too combined to me. Why don't we use the SECT_4K flag and another flag to indicate it's a PMC part? Even better, I recall you can just read the chip jedec ID and determine if it's a PMC part according to that. Then if it is PMC AND the SECT_4K flag is set, there is no need to add another flag at all, no? > But in the meantime, I fixed a conflict with stuff that's been applied > since you submitted this patch, and applied to l2-mtd.git. Thanks! > > Brian > > On Mon, Jul 15, 2013 at 12:13:56PM +0200, Michel Stempin wrote: > > Add support for PMC (now Chingis, part of ISSI) Pm25LV512 (512 kBbit), > > Pm25LV010 (1 Mbit) and Pm25LQ032 (32 Mbit) SPI Flash chips. > > > > This patch addresses two generations of PMC SPI Flash chips: > > - Pm25LV512 and Pm25LV010: these have 4KB sectors and 32KB > > > > blocks. The 4KB sector erase uses a non-standard opcode > > (0xd7). They do not support JEDEC RDID (0x9f), and so they can only > > be detected by matching their name string with pre-configured > > platform data. Because of the cascaded acquisitions, the datasheet > > is no longer available on the current manufacturer's website, > > although it is still commonly used in some recent wireless routers > > (<https://forum.openwrt.org/viewtopic.php?pid=186360#p186360>). The > > only public datasheet available seems to be on GeoCities: > > <http://www.geocities.jp/scottle556/pdf/Pm25LV512-010.pdf> > > > > - Pm25LQ032: a newer generation flash, with 4KB sectors and 32KB > > > > blocks. It uses the standard erase and JEDEC read-ID > > opcodes. Manufacturer's datasheet is here: > > <http://www.chingistek.com/img/Product_Files/Pm25LQ032C%20datasheet%20 > > v1.6.1.pdf> > > > > This patch is resent in order to take into account both Brian Norris > > remarks and this upstream patch: > > > > commit e534ee4f9ca29fdb38eea4b0c53f2154fbd8c1ee > > Author: Krzysztof Mazur <krzysiek@podlesie.net> > > Date: Fri Feb 22 15:51:05 2013 +0100 > > > > mtd: m25p80: introduce SST_WRITE flag for SST byte programming > > > > Not all SST devices implement the SST byte programming command. > > Some devices (like SST25VF064C) implement only standard m25p80 page > > write command. > > > > Now SPI flash devices that need sst_write() are explicitly marked > > with new SST_WRITE flag and the decision to use sst_write() is based > > on this flag instead of manufacturer id. > > > > Signed-off-by: Krzysztof Mazur <krzysiek@podlesie.net> > > Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> > > > > Signed-off-by: Gabor Juhos <juhosg@openwrt.org> > > Signed-off-by: Michel Stempin <michel.stempin@wanadoo.fr> > > --- > > > > drivers/mtd/devices/m25p80.c | 10 ++++++++++ > > 1 file changed, 10 insertions(+) > > > > diff --git a/drivers/mtd/devices/m25p80.c b/drivers/mtd/devices/m25p80.c > > index 2f3d2a5..55c8b5f 100644 > > --- a/drivers/mtd/devices/m25p80.c > > +++ b/drivers/mtd/devices/m25p80.c > > @@ -43,6 +43,7 @@ > > > > #define OPCODE_FAST_READ 0x0b /* Read data bytes (high frequency) */ > > #define OPCODE_PP 0x02 /* Page program (up to 256 bytes) */ > > #define OPCODE_BE_4K 0x20 /* Erase 4KiB block */ > > > > +#define OPCODE_BE_4K_PMC 0xd7 /* Erase 4KiB block on PMC chips */ > > > > #define OPCODE_BE_32K 0x52 /* Erase 32KiB block */ > > #define OPCODE_CHIP_ERASE 0xc7 /* Erase whole flash chip */ > > #define OPCODE_SE 0xd8 /* Sector erase (usually 64KiB) */ > > > > @@ -682,6 +683,7 @@ struct flash_info { > > > > #define SECT_4K 0x01 /* OPCODE_BE_4K works uniformly */ > > #define M25P_NO_ERASE 0x02 /* No erase command needed */ > > #define SST_WRITE 0x04 /* use SST byte programming */ > > > > +#define SECT_4K_PMC 0x08 /* OPCODE_BE_4K_PMC works uniformly */ > > > > }; > > > > #define INFO(_jedec_id, _ext_id, _sector_size, _n_sectors, _flags) \ > > > > @@ -762,6 +764,11 @@ static const struct spi_device_id m25p_ids[] = { > > > > { "n25q128a13", INFO(0x20ba18, 0, 64 * 1024, 256, 0) }, > > { "n25q256a", INFO(0x20ba19, 0, 64 * 1024, 512, SECT_4K) }, > > > > + /* PMC */ > > + { "pm25lv512", INFO(0, 0, 32 * 1024, 2, SECT_4K_PMC) }, > > + { "pm25lv010", INFO(0, 0, 32 * 1024, 4, SECT_4K_PMC) }, > > + { "pm25lq032", INFO(0x7f9d46, 0, 64 * 1024, 64, SECT_4K) }, > > + > > > > /* Spansion -- single (large) sector size only, at least > > > > * for the chips listed here (without boot sectors). > > */ > > > > @@ -1014,6 +1021,9 @@ static int m25p_probe(struct spi_device *spi) > > > > if (info->flags & SECT_4K) { > > > > flash->erase_opcode = OPCODE_BE_4K; > > flash->mtd.erasesize = 4096; > > > > + } else if (info->flags & SECT_4K_PMC) { > > + flash->erase_opcode = OPCODE_BE_4K_PMC; > > + flash->mtd.erasesize = 4096; > > > > } else { > > > > flash->erase_opcode = OPCODE_SE; > > flash->mtd.erasesize = info->sector_size; Best regards, Marek Vasut
On Wed, Aug 21, 2013 at 09:41:38AM +0200, Marek Vasut wrote: > Dear Brian Norris, > > > + Marek, since he's been reviewing (with dismay?) the increase in macro > > flags in this driver. If there are any objections, I can amend/drop the > > patch. > > Hmmm ... this SECT_4K_PMC seems too combined to me. Why don't we use the SECT_4K > flag and another flag to indicate it's a PMC part? Even better, I recall you can Separating manufacturer from SECT_4K sounds good, but it really doesn't buy us much. See my next comments. > just read the chip jedec ID and determine if it's a PMC part according to that. > Then if it is PMC AND the SECT_4K flag is set, there is no need to add another > flag at all, no? IIUC, Michel's comment applies: "They do not support JEDEC RDID (0x9f), and so they can only be detected by matching their name string with pre-configured platform data." So we cannot use RDID to identify by manufacturer. In fact, this same point screws up any attempt at manufacturer-based property detection for non-JEDEC devices. I guess we just can't expect much from such devices. So we would have to introduce two flags to the table: one to flag the manufacturer and one to flag the opcode. Not necessary, IMO. > > On Mon, Jul 15, 2013 at 12:13:56PM +0200, Michel Stempin wrote: [...] > > > @@ -762,6 +764,11 @@ static const struct spi_device_id m25p_ids[] = { > > > > > > { "n25q128a13", INFO(0x20ba18, 0, 64 * 1024, 256, 0) }, > > > { "n25q256a", INFO(0x20ba19, 0, 64 * 1024, 512, SECT_4K) }, > > > > > > + /* PMC */ > > > + { "pm25lv512", INFO(0, 0, 32 * 1024, 2, SECT_4K_PMC) }, > > > + { "pm25lv010", INFO(0, 0, 32 * 1024, 4, SECT_4K_PMC) }, Note that only the non-JEDEC chips needed the old commands. > > > + { "pm25lq032", INFO(0x7f9d46, 0, 64 * 1024, 64, SECT_4K) }, > > > + > > > > > > /* Spansion -- single (large) sector size only, at least > > > > > > * for the chips listed here (without boot sectors). > > > */ > > > Brian
Dear Brian Norris, > On Wed, Aug 21, 2013 at 09:41:38AM +0200, Marek Vasut wrote: > > Dear Brian Norris, > > > > > + Marek, since he's been reviewing (with dismay?) the increase in macro > > > flags in this driver. If there are any objections, I can amend/drop the > > > patch. > > > > Hmmm ... this SECT_4K_PMC seems too combined to me. Why don't we use the > > SECT_4K flag and another flag to indicate it's a PMC part? Even better, > > I recall you can > > Separating manufacturer from SECT_4K sounds good, but it really doesn't > buy us much. See my next comments. I see, that's really bad news. Thanks for the explanation! I guess there really is nothing much we can do about such parts. But then if we take device tree probe into consideration, we might actually want to match the part name to discern the PMS device. Or am I talking complete nonsense? Best regards, Marek Vasut
On Wed, Aug 21, 2013 at 10:07:17AM +0200, Marek Vasut wrote: > > On Wed, Aug 21, 2013 at 09:41:38AM +0200, Marek Vasut wrote: > > > > + Marek, since he's been reviewing (with dismay?) the increase in macro > > > > flags in this driver. If there are any objections, I can amend/drop the > > > > patch. > > > > > > Hmmm ... this SECT_4K_PMC seems too combined to me. Why don't we use the > > > SECT_4K flag and another flag to indicate it's a PMC part? Even better, > > > I recall you can > > > > Separating manufacturer from SECT_4K sounds good, but it really doesn't > > buy us much. See my next comments. > > I see, that's really bad news. Thanks for the explanation! > > I guess there really is nothing much we can do about such parts. But then if we > take device tree probe into consideration, we might actually want to match the > part name to discern the PMS device. Or am I talking complete nonsense? I don't think the device tree probe really gives us anything different than the platform_device probe (a non-JEDEC device can be matched via device-tree "compatible" property or via platform_device "name" property, I think?). So in either case, are you suggesting a string comparison for "pm25" on the spi_device_id.name field? Seems a bit like nonsense :) Additionally, this still doesn't solve the problem that the old PMC chips need the special opcode, but the newer one doesn't. Brian
Hi Brian, > On Wed, Aug 21, 2013 at 10:07:17AM +0200, Marek Vasut wrote: > > > On Wed, Aug 21, 2013 at 09:41:38AM +0200, Marek Vasut wrote: > > > > > + Marek, since he's been reviewing (with dismay?) the increase in > > > > > macro flags in this driver. If there are any objections, I can > > > > > amend/drop the patch. > > > > > > > > Hmmm ... this SECT_4K_PMC seems too combined to me. Why don't we use > > > > the SECT_4K flag and another flag to indicate it's a PMC part? Even > > > > better, I recall you can > > > > > > Separating manufacturer from SECT_4K sounds good, but it really doesn't > > > buy us much. See my next comments. > > > > I see, that's really bad news. Thanks for the explanation! > > > > I guess there really is nothing much we can do about such parts. But then > > if we take device tree probe into consideration, we might actually want > > to match the part name to discern the PMS device. Or am I talking > > complete nonsense? > > I don't think the device tree probe really gives us anything different > than the platform_device probe (a non-JEDEC device can be matched via > device-tree "compatible" property or via platform_device "name" > property, I think?). So in either case, are you suggesting a string > comparison for "pm25" on the spi_device_id.name field? Seems a bit > like nonsense :) Yeah, you're right. > Additionally, this still doesn't solve the problem that the old PMC > chips need the special opcode, but the newer one doesn't. One would have to match the full part name, but that's already happening. Please ignore me, I had my coffee only now. Best regards, Marek Vasut
On Wed, Aug 21, 2013 at 6:10 AM, Marek Vasut <marex@denx.de> wrote: >> On Wed, Aug 21, 2013 at 10:07:17AM +0200, Marek Vasut wrote: >> > > On Wed, Aug 21, 2013 at 09:41:38AM +0200, Marek Vasut wrote: >> > > > > + Marek, since he's been reviewing (with dismay?) the increase in >> > > > > macro flags in this driver. If there are any objections, I can >> > > > > amend/drop the patch. >> > > > >> > > > Hmmm ... this SECT_4K_PMC seems too combined to me. Why don't we use >> > > > the SECT_4K flag and another flag to indicate it's a PMC part? Even >> > > > better, I recall you can >> > > >> > > Separating manufacturer from SECT_4K sounds good, but it really doesn't >> > > buy us much. See my next comments. >> > >> > I see, that's really bad news. Thanks for the explanation! >> > >> > I guess there really is nothing much we can do about such parts. But then >> > if we take device tree probe into consideration, we might actually want >> > to match the part name to discern the PMS device. Or am I talking >> > complete nonsense? >> >> I don't think the device tree probe really gives us anything different >> than the platform_device probe (a non-JEDEC device can be matched via >> device-tree "compatible" property or via platform_device "name" >> property, I think?). So in either case, are you suggesting a string >> comparison for "pm25" on the spi_device_id.name field? Seems a bit >> like nonsense :) > > Yeah, you're right. > >> Additionally, this still doesn't solve the problem that the old PMC >> chips need the special opcode, but the newer one doesn't. > > One would have to match the full part name, but that's already happening. Please > ignore me, I had my coffee only now. No problem. Then I'm leaving the patch queued as-is. Thanks for taking a look. Brian
diff --git a/drivers/mtd/devices/m25p80.c b/drivers/mtd/devices/m25p80.c index 2f3d2a5..55c8b5f 100644 --- a/drivers/mtd/devices/m25p80.c +++ b/drivers/mtd/devices/m25p80.c @@ -43,6 +43,7 @@ #define OPCODE_FAST_READ 0x0b /* Read data bytes (high frequency) */ #define OPCODE_PP 0x02 /* Page program (up to 256 bytes) */ #define OPCODE_BE_4K 0x20 /* Erase 4KiB block */ +#define OPCODE_BE_4K_PMC 0xd7 /* Erase 4KiB block on PMC chips */ #define OPCODE_BE_32K 0x52 /* Erase 32KiB block */ #define OPCODE_CHIP_ERASE 0xc7 /* Erase whole flash chip */ #define OPCODE_SE 0xd8 /* Sector erase (usually 64KiB) */ @@ -682,6 +683,7 @@ struct flash_info { #define SECT_4K 0x01 /* OPCODE_BE_4K works uniformly */ #define M25P_NO_ERASE 0x02 /* No erase command needed */ #define SST_WRITE 0x04 /* use SST byte programming */ +#define SECT_4K_PMC 0x08 /* OPCODE_BE_4K_PMC works uniformly */ }; #define INFO(_jedec_id, _ext_id, _sector_size, _n_sectors, _flags) \ @@ -762,6 +764,11 @@ static const struct spi_device_id m25p_ids[] = { { "n25q128a13", INFO(0x20ba18, 0, 64 * 1024, 256, 0) }, { "n25q256a", INFO(0x20ba19, 0, 64 * 1024, 512, SECT_4K) }, + /* PMC */ + { "pm25lv512", INFO(0, 0, 32 * 1024, 2, SECT_4K_PMC) }, + { "pm25lv010", INFO(0, 0, 32 * 1024, 4, SECT_4K_PMC) }, + { "pm25lq032", INFO(0x7f9d46, 0, 64 * 1024, 64, SECT_4K) }, + /* Spansion -- single (large) sector size only, at least * for the chips listed here (without boot sectors). */ @@ -1014,6 +1021,9 @@ static int m25p_probe(struct spi_device *spi) if (info->flags & SECT_4K) { flash->erase_opcode = OPCODE_BE_4K; flash->mtd.erasesize = 4096; + } else if (info->flags & SECT_4K_PMC) { + flash->erase_opcode = OPCODE_BE_4K_PMC; + flash->mtd.erasesize = 4096; } else { flash->erase_opcode = OPCODE_SE; flash->mtd.erasesize = info->sector_size;