After I added SFDP 1.6 parser I found in Micron's datasheet for MT25Q 512Mb chip (MT25QL512AB/MT25QU512AB) that the chip returnes SFDP revision 1.5 but simultaniously provides the required information about addressing mode switch, however all information about direct-4BA functions isn't provided.
I decided to parse new SFDP 4-bytes info from SFDP revision 1.5 besides of direct-4BA instructions. This assumption is made for better support of Micron flash chips.
This code should be checked and may be modified or removed.
Patched files ------------- sfdp.c + added fix for Micron chips to read 16th dword stating from SFDP revision 1.5
Signed-off-by: Boris Baykov dev@borisbaykov.com, Russia, Jan 2014 --- diff -U 5 -N ./flashrom/sfdp.c ./flashrom-1868-4ba-9/sfdp.c --- ./flashrom/sfdp.c 2015-01-11 01:55:16.000000000 +0300 +++ ./flashrom-1868-4ba-9/sfdp.c 2015-01-10 21:33:06.000000000 +0300 @@ -19,10 +19,11 @@ */
/* * History of changes: * 05/01/2014 Added compliance to JESD216B standard and SFDP revision 1.6 + * 07/01/2014 Modified to support SFDP revision 1.5 (for Micron flash chips) */
#include <stdint.h> #include <stdlib.h> #include <string.h> @@ -177,11 +178,11 @@ __func__, eraser_type); return 1; }
/* Parse of JEDEC SFDP Basic Flash Parameter Table */ -static int sfdp_fill_flash(struct flashchip *chip, uint8_t *buf, uint16_t len, int sfdp_rev_16) +static int sfdp_fill_flash(struct flashchip *chip, uint8_t *buf, uint16_t len, int sfdp_rev_15) { uint8_t opcode_4k_erase = 0xFF; uint32_t tmp32; uint8_t tmp8; uint32_t total_size; /* in bytes */ @@ -296,14 +297,14 @@ total_size = ((tmp32 & 0x7FFFFFFF) + 1) / 8; chip->total_size = total_size / 1024; msg_cdbg2(" Flash chip size is %d kB.\n", chip->total_size);
if (total_size > (1 << 24)) { - if(!sfdp_rev_16) { + if(!sfdp_rev_15) { msg_cdbg("Flash chip size is bigger than what 3-Byte addressing " "can access but chip's SFDP revision is lower than 1.6 " - "\nConsequently 4-bytes addressing can NOT be " + "(1.5).\nConsequently 4-bytes addressing can NOT be " "properly configured using current SFDP information.\n"); #ifndef FBA_USE_EXT_ADDR_REG_BY_DEFAULT msg_cdbg("Assuming that 4-bytes addressing mode can be " "entered by CMD B7h with WREN.\n"); #else @@ -374,14 +375,14 @@ because we don't have to know how to switch to 4-bytes mode and back when we are already in 4-bytes mode permanently. */ if (chip->feature_bits & FEATURE_4BA_ONLY) goto done;
- /* If the SFDP revision supported by the chip is lower that 1.6 + /* If the SFDP revision supported by the chip is lower that 1.6 (1.5) we can not read and analyze 16th DWORD of Basic Flash Parameter Table. Using defaults by FBA_USE_EXT_ADDR_REG_BY_DEFAULT define. */ - if(!sfdp_rev_16) + if(!sfdp_rev_15) goto done;
if (len < 16 * 4) { msg_cdbg("%s: len of BFPT is out of spec\n", __func__); msg_cerr("ERROR: Unable read 4-bytes addressing parameters.\n"); @@ -600,11 +601,11 @@ /* need to limit the table loop by comparing i to uint8_t nph hence: */ uint16_t i; struct sfdp_tbl_hdr *hdrs; uint8_t *hbuf; uint8_t *tbuf; - int sfdp_rev_16 = 0; + int sfdp_rev_16 = 0, sfdp_rev_15 = 0;
if (spi_sfdp_read_sfdp(flash, 0x00, buf, 4)) { msg_cdbg("Receiving SFDP signature failed.\n"); return 0; } @@ -634,12 +635,29 @@ /* JEDEC JESD216B defines SFDP revision 1.6 and includes: 1) 16 dwords in Basic Flash Parameter Table 2) 16th dword has information how to enter and exit 4-bytes addressing mode 3) 4-Bytes Address Instruction Table with ID 0xFF84 + + However we can see in the datasheet for Micron's + MT25Q 512Mb chip (MT25QL512AB/MT25QU512AB) that the + chip returnes SFDP revision 1.5 and has 16 dwords + in its Basic Flash Paramater Table. Also the information + about addressing mode switch is exist in the 16th dword. + But 4-Bytes Address Instruction Table is absent. + + So we will use 16th dword from SFDP revision 1.5 + but 4-Bytes Address Instruction Table from SFDP 1.6 only. + This assumption is made for better support of Micron + flash chips. + + FIXME: SFDP revisions compliance should be checked + more carefully after more information about JESD216B + SFDP tables will be known from real flash chips. */ sfdp_rev_16 = (buf[1] == 1 && buf[0] >= 6) || buf[1] > 1; + sfdp_rev_15 = (buf[1] == 1 && buf[0] >= 5) || buf[1] > 1;
nph = buf[2]; msg_cdbg2("SFDP number of parameter headers is %d (NPH = %d).\n", nph + 1, nph);
@@ -715,21 +733,21 @@ "parameter table is not 0xFF00 as" "demanded by JESD216 (warning only)." "\n"); #ifdef JESD216B_SIMULATION if(!sfdp_jesd216b_simulation_dw16(&tbuf, &len)) - sfdp_rev_16 = 1; /* pretend as SFDP rev 1.6 */ + sfdp_rev_16 = sfdp_rev_15 = 1; /* pretend as SFDP rev 1.6 */ #endif if (hdrs[i].v_major != 0x01) { msg_cdbg("The chip contains an unknown " "version of the JEDEC flash " "parameters table, skipping it.\n"); } else if (len != 16 * 4 && len != 9 * 4 && len != 4 * 4) { msg_cdbg("Length of the mandatory JEDEC SFDP " "parameter table is wrong (%d B), " "skipping it.\n", len); - } else if (sfdp_fill_flash(flash->chip, tbuf, len, sfdp_rev_16) == 0) + } else if (sfdp_fill_flash(flash->chip, tbuf, len, sfdp_rev_15) == 0) ret = 1; #ifdef JESD216B_SIMULATION if(ret == 1 && !sfdp_jesd216b_simulation_4bait(&tbuf, &len)) sfdp_parse_4ba_table(flash->chip, tbuf, len); #endif
* Date of changes fixed from 2014 to 2015. It's correct now.
The corrected file of [PATCH 9/10] is here below.
After I added SFDP 1.6 parser I found in Micron's datasheet for MT25Q 512Mb chip (MT25QL512AB/MT25QU512AB) that the chip returnes SFDP revision 1.5 but simultaniously provides the required information about addressing mode switch, however all information about direct-4BA functions isn't provided.
I decided to parse new SFDP 4-bytes info from SFDP revision 1.5 besides of direct-4BA instructions. This assumption is made for better support of Micron flash chips.
This code should be checked and may be modified or removed.
Patched files ------------- sfdp.c + added fix for Micron chips to read 16th dword stating from SFDP revision 1.5
Signed-off-by: Boris Baykov dev@borisbaykov.com, Russia, Jan 2014 --- diff -U 5 -N ./flashrom/sfdp.c ./flashrom-1868-4ba-9/sfdp.c --- ./flashrom/sfdp.c 2015-01-16 17:06:21.000000000 +0300 +++ ./flashrom-1868-4ba-9/sfdp.c 2015-01-16 17:01:42.000000000 +0300 @@ -19,10 +19,11 @@ */
/* * History of changes: * 05/01/2015 Added compliance to JESD216B standard and SFDP revision 1.6 + * 07/01/2015 Modified to support SFDP revision 1.5 (for Micron flash chips) */
#include <stdint.h> #include <stdlib.h> #include <string.h> @@ -177,11 +178,11 @@ __func__, eraser_type); return 1; }
/* Parse of JEDEC SFDP Basic Flash Parameter Table */ -static int sfdp_fill_flash(struct flashchip *chip, uint8_t *buf, uint16_t len, int sfdp_rev_16) +static int sfdp_fill_flash(struct flashchip *chip, uint8_t *buf, uint16_t len, int sfdp_rev_15) { uint8_t opcode_4k_erase = 0xFF; uint32_t tmp32; uint8_t tmp8; uint32_t total_size; /* in bytes */ @@ -296,14 +297,14 @@ total_size = ((tmp32 & 0x7FFFFFFF) + 1) / 8; chip->total_size = total_size / 1024; msg_cdbg2(" Flash chip size is %d kB.\n", chip->total_size);
if (total_size > (1 << 24)) { - if(!sfdp_rev_16) { + if(!sfdp_rev_15) { msg_cdbg("Flash chip size is bigger than what 3-Byte addressing " "can access but chip's SFDP revision is lower than 1.6 " - "\nConsequently 4-bytes addressing can NOT be " + "(1.5).\nConsequently 4-bytes addressing can NOT be " "properly configured using current SFDP information.\n"); #ifndef FBA_USE_EXT_ADDR_REG_BY_DEFAULT msg_cdbg("Assuming that 4-bytes addressing mode can be " "entered by CMD B7h with WREN.\n"); #else @@ -374,14 +375,14 @@ because we don't have to know how to switch to 4-bytes mode and back when we are already in 4-bytes mode permanently. */ if (chip->feature_bits & FEATURE_4BA_ONLY) goto done;
- /* If the SFDP revision supported by the chip is lower that 1.6 + /* If the SFDP revision supported by the chip is lower that 1.6 (1.5) we can not read and analyze 16th DWORD of Basic Flash Parameter Table. Using defaults by FBA_USE_EXT_ADDR_REG_BY_DEFAULT define. */ - if(!sfdp_rev_16) + if(!sfdp_rev_15) goto done;
if (len < 16 * 4) { msg_cdbg("%s: len of BFPT is out of spec\n", __func__); msg_cerr("ERROR: Unable read 4-bytes addressing parameters.\n"); @@ -600,11 +601,11 @@ /* need to limit the table loop by comparing i to uint8_t nph hence: */ uint16_t i; struct sfdp_tbl_hdr *hdrs; uint8_t *hbuf; uint8_t *tbuf; - int sfdp_rev_16 = 0; + int sfdp_rev_16 = 0, sfdp_rev_15 = 0;
if (spi_sfdp_read_sfdp(flash, 0x00, buf, 4)) { msg_cdbg("Receiving SFDP signature failed.\n"); return 0; } @@ -634,12 +635,29 @@ /* JEDEC JESD216B defines SFDP revision 1.6 and includes: 1) 16 dwords in Basic Flash Parameter Table 2) 16th dword has information how to enter and exit 4-bytes addressing mode 3) 4-Bytes Address Instruction Table with ID 0xFF84 + + However we can see in the datasheet for Micron's + MT25Q 512Mb chip (MT25QL512AB/MT25QU512AB) that the + chip returnes SFDP revision 1.5 and has 16 dwords + in its Basic Flash Paramater Table. Also the information + about addressing mode switch is exist in the 16th dword. + But 4-Bytes Address Instruction Table is absent. + + So we will use 16th dword from SFDP revision 1.5 + but 4-Bytes Address Instruction Table from SFDP 1.6 only. + This assumption is made for better support of Micron + flash chips. + + FIXME: SFDP revisions compliance should be checked + more carefully after more information about JESD216B + SFDP tables will be known from real flash chips. */ sfdp_rev_16 = (buf[1] == 1 && buf[0] >= 6) || buf[1] > 1; + sfdp_rev_15 = (buf[1] == 1 && buf[0] >= 5) || buf[1] > 1;
nph = buf[2]; msg_cdbg2("SFDP number of parameter headers is %d (NPH = %d).\n", nph + 1, nph);
@@ -715,21 +733,21 @@ "parameter table is not 0xFF00 as" "demanded by JESD216 (warning only)." "\n"); #ifdef JESD216B_SIMULATION if(!sfdp_jesd216b_simulation_dw16(&tbuf, &len)) - sfdp_rev_16 = 1; /* pretend as SFDP rev 1.6 */ + sfdp_rev_16 = sfdp_rev_15 = 1; /* pretend as SFDP rev 1.6 */ #endif if (hdrs[i].v_major != 0x01) { msg_cdbg("The chip contains an unknown " "version of the JEDEC flash " "parameters table, skipping it.\n"); } else if (len != 16 * 4 && len != 9 * 4 && len != 4 * 4) { msg_cdbg("Length of the mandatory JEDEC SFDP " "parameter table is wrong (%d B), " "skipping it.\n", len); - } else if (sfdp_fill_flash(flash->chip, tbuf, len, sfdp_rev_16) == 0) + } else if (sfdp_fill_flash(flash->chip, tbuf, len, sfdp_rev_15) == 0) ret = 1; #ifdef JESD216B_SIMULATION if(ret == 1 && !sfdp_jesd216b_simulation_4bait(&tbuf, &len)) sfdp_parse_4ba_table(flash->chip, tbuf, len); #endif
-
Boris Baykov