OneNAND: Remove base address offset usage
RFC PATCH 1/1 OneNAND: Remove base address offset usage
While locally preparing some U-Boot patches for ARM based OMAP3 boards, some using OneNAND and some using NAND, we found some differences in OneNAND and NAND command address handling.
As this might confuse users (it already confused us), we like to align OneNAND and NAND address handling.
The issue is that cmd_onenand.c subtracts the onenand base address from the addresses you type into the u-boot command line so, unlike nand, you can't use addresses relative to the start of the onenand part e.g. this won't work:
onenand read 82000000 280000 400000
you have to use:
onenand read 82000000 20280000 400000
Looking at recent git, the only board currently using OneNAND is Apollon, and for this the OneNAND base address is 0 (apollon.h)
#define CFG_ONENAND_BASE 0x00000000
so patch below won't break any existing boards and will align OneNAND and NAND handling on boards where OneNAND base address is != 0.
---
common/cmd_onenand.c +0 -6
@@ -58,8 +58,6 @@ int do_onenand(cmd_tbl_t * cmdtp, int fl }else { start =simple_strtoul (argv[2 ],NULL ,10 ); end =simple_strtoul (argv[3 ],NULL ,10 ); - start -= (unsigned long )onenand_chip.base; - end -= (unsigned long )onenand_chip.base; start >>= onenand_chip.erase_shift; end >>= onenand_chip.erase_shift;
@@ -92,8 +90,6 @@ int do_onenand(cmd_tbl_t * cmdtp, int flsize_t retlen =0 ;int oob =strncmp (argv[1 ],"read.oob" ,8 ) ?0 :1 ; - ofs -= (unsigned long )onenand_chip.base; -if (oob)onenand_read_oob (&onenand_mtd, ofs, len, &retlen, (u_char *) addr);
@@ -111,8 +107,6 @@ int do_onenand(cmd_tbl_t * cmdtp, int flsize_t len =simple_strtoul (argv[4 ],NULL ,16 );size_t retlen =0 ; - ofs -= (unsigned long )onenand_chip.base; -onenand_write (&onenand_mtd, ofs, len, &retlen, (u_char *) addr);printf ("Done\n" );
common/env_onenand.c +0 -3
@@ -66,7 +66,6 @@ void env_relocate_spec(void)size_t retlen; env_addr =CFG_ENV_ADDR ; - env_addr -= (unsigned long ) onenand_chip.base;/* Check OneNAND exist */ if (onenand_mtd.oobblock)
@@ -101,7 +100,6 @@ int saveenv(void) instr.len =CFG_ENV_SIZE ; instr.addr = env_addr; - instr.addr -= (unsigned long )onenand_chip.base;if (onenand_erase (&onenand_mtd, &instr)) {printf ("OneNAND: erase failed at 0x%08lx\n" , env_addr);return 1 ;
@@ -111,7 +109,6 @@ int saveenv(void) env_ptr->crc =crc32 (0 , env_ptr->data,ONENAND_ENV_SIZE (onenand_mtd)); - env_addr -= (unsigned long )onenand_chip.base;if (onenand_write (&onenand_mtd, env_addr, onenand_mtd.oobblock, &retlen, (u_char *) env_ptr)) {printf ("OneNAND: write failed at 0x%08x\n" , instr.addr);
Wolfgang Denk
AppliedRFC PATCH 1/1 OneNAND: Remove base address offset usage
Dear Dirk,
Quoted 20 lines · In message <[email protected]> you wrote:
While locally preparing some U-Boot patches for ARM based OMAP3 boards, some using OneNAND and some using NAND, we found some differences in OneNAND and NAND command address handling.
As this might confuse users (it already confused us), we like to align OneNAND and NAND address handling.
The issue is that
cmd_onenand.csubtracts the onenand base address from the addresses you type into the u-boot command line so, unlike nand, you can't use addresses relative to the start of the onenand part e.g. this won't work:onenand read 82000000 280000 400000
you have to use:
onenand read 82000000 20280000 400000
Looking at recent git, the only board currently using OneNAND is Apollon, and for this the OneNAND base address is 0 (
apollon.h)
Quoted 8 lines
#define
CFG_ONENAND_BASE0x00000000so patch below won't break any existing boards and will align OneNAND and NAND handling on boards where OneNAND base address is != 0.
Signed-off-by Steve Sakoman <[email protected]>Signed-off-by Manikandan Pillai <[email protected]>Signed-off-by Dirk Behme <[email protected]>
Applied, thanks.
Best regards,
Wolfgang Denk
Signature · 5 lines
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de Crash programs fail because they are based on the theory that, with nine women pregnant, you can get a baby a month. - Wernher von Braun