| Message ID | 1304502284-392-1-git-send-email-sr@denx.de |
|---|---|
| State | Accepted |
| Commit | a89a990159d450765b8a5675e640525bc6258d1b |
| Headers | show |
On Wed, 4 May 2011 11:44:44 +0200 Stefan Roese <sr@denx.de> wrote: > Patch 65a9db7b [nand_spl: Fix large page nand_command()] broke > nand booting on canyonlands. "options" has to be initialized to > 0. If not, boards might have the NAND_BUSWIDTH_16 bit set, > resulting in wrong offset calculation. > > Signed-off-by: Stefan Roese <sr@denx.de> > Cc: Scott Wood <scottwood@freescale.com> > Cc: Alex Waterman <awaterman@dawning.com> > --- > nand_spl/nand_boot.c | 1 + > 1 files changed, 1 insertions(+), 0 deletions(-) > > diff --git a/nand_spl/nand_boot.c b/nand_spl/nand_boot.c > index cefeb91..62303aa 100644 > --- a/nand_spl/nand_boot.c > +++ b/nand_spl/nand_boot.c > @@ -232,6 +232,7 @@ void nand_boot(void) > nand_info.priv = &nand_chip; > nand_chip.IO_ADDR_R = nand_chip.IO_ADDR_W = (void __iomem *)CONFIG_SYS_NAND_BASE; > nand_chip.dev_ready = NULL; /* preset to NULL */ > + nand_chip.options = 0; > board_nand_init(&nand_chip); > > if (nand_chip.select_chip) Maybe nand_chip should be made static so it goes in the BSS? This would avoid any other such problems, without adding any extra code to zero things out. -Scott
On Wednesday 04 May 2011 19:30:29 Scott Wood wrote: > > + nand_chip.options = 0; > > board_nand_init(&nand_chip); > > > > if (nand_chip.select_chip) > > Maybe nand_chip should be made static so it goes in the BSS? This would > avoid any other such problems, without adding any extra code to zero things > out. Yes, good idea. I tried it but it resulted in a bigger image instead of a smaller one. I need to investigate further to see why this is the case... Cheers, Stefan -- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-0 Fax: (+49)-8142-66989-80 Email: office@denx.de
On Wed, May 04, 2011 at 11:44:44AM +0200, Stefan Roese wrote: > Patch 65a9db7b [nand_spl: Fix large page nand_command()] broke > nand booting on canyonlands. "options" has to be initialized to > 0. If not, boards might have the NAND_BUSWIDTH_16 bit set, > resulting in wrong offset calculation. > > Signed-off-by: Stefan Roese <sr@denx.de> > Cc: Scott Wood <scottwood@freescale.com> > Cc: Alex Waterman <awaterman@dawning.com> > --- > nand_spl/nand_boot.c | 1 + > 1 files changed, 1 insertions(+), 0 deletions(-) Applied to u-boot-nand-flash -Scott
diff --git a/nand_spl/nand_boot.c b/nand_spl/nand_boot.c index cefeb91..62303aa 100644 --- a/nand_spl/nand_boot.c +++ b/nand_spl/nand_boot.c @@ -232,6 +232,7 @@ void nand_boot(void) nand_info.priv = &nand_chip; nand_chip.IO_ADDR_R = nand_chip.IO_ADDR_W = (void __iomem *)CONFIG_SYS_NAND_BASE; nand_chip.dev_ready = NULL; /* preset to NULL */ + nand_chip.options = 0; board_nand_init(&nand_chip); if (nand_chip.select_chip)
Patch 65a9db7b [nand_spl: Fix large page nand_command()] broke nand booting on canyonlands. "options" has to be initialized to 0. If not, boards might have the NAND_BUSWIDTH_16 bit set, resulting in wrong offset calculation. Signed-off-by: Stefan Roese <sr@denx.de> Cc: Scott Wood <scottwood@freescale.com> Cc: Alex Waterman <awaterman@dawning.com> --- nand_spl/nand_boot.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-)