[U-Boot] [PATCH v2 1/5] spi, spi_mxc: do not hang in spi_xchg_single

[Jagan Teki]

On Sat, Jul 12, 2014 at 9:40 AM, Heiko Schocher <hs at denx.de> wrote:
> if status register do never set MXC_CSPICTRL_TC, spi_xchg_single
> endless loops. Add a timeout here to prevent endless hang.
>
> Signed-off-by: Heiko Schocher <hs at denx.de>
> Cc: Dirk Behme <dirk.behme at gmail.com>
> Cc: Jagannadha Sutradharudu Teki <jagannadh.teki at gmail.com>
>
> ---
> - changes for v2:
>   - use timer api to poll till TC bit is set as Jagan Teki suggested
>     and make this timeout configurable through CONFIG_SYS_SPI_MXC_WAIT
>
> - repost this patch in this series again, orginal post:
> Patchwork [U-Boot,v2,02/04] spi, spi_mxc: do not hang in spi_xchg_single
> http://patchwork.ozlabs.org/patch/354768/
>
>  README                |  4 ++++
>  drivers/spi/mxc_spi.c | 18 ++++++++++++++++--
>  2 files changed, 20 insertions(+), 2 deletions(-)
>
> diff --git a/README b/README
> index 4ac7399..0f24bd3 100644
> --- a/README
> +++ b/README
> @@ -2597,6 +2597,10 @@ CBFS (Coreboot Filesystem) support
>                 Enables the driver for the SPI controllers on i.MX and MXC
>                 SoCs. Currently i.MX31/35/51 are supported.
>
> +               CONFIG_SYS_SPI_MXC_WAIT
> +               Timeout for waiting until spi transfer completed.
> +               default: (CONFIG_SYS_HZ/100)     /* 10 ms */
> +
>  - FPGA Support: CONFIG_FPGA
>
>                 Enables FPGA subsystem.
> diff --git a/drivers/spi/mxc_spi.c b/drivers/spi/mxc_spi.c
> index f3f029d..4732850 100644
> --- a/drivers/spi/mxc_spi.c
> +++ b/drivers/spi/mxc_spi.c
> @@ -30,6 +30,10 @@ static unsigned long spi_bases[] = {
>  #define reg_read readl
>  #define reg_write(a, v) writel(v, a)
>
> +#if !defined(CONFIG_SYS_SPI_MXC_WAIT)
> +#define CONFIG_SYS_SPI_MXC_WAIT                (CONFIG_SYS_HZ/100)     /* 10 ms */
> +#endif
> +
>  struct mxc_spi_slave {
>         struct spi_slave slave;
>         unsigned long   base;
> @@ -212,6 +216,8 @@ int spi_xchg_single(struct spi_slave *slave, unsigned int bitlen,
>         int nbytes = DIV_ROUND_UP(bitlen, 8);
>         u32 data, cnt, i;
>         struct cspi_regs *regs = (struct cspi_regs *)mxcs->base;
> +       u32 ts;
> +       int status;
>
>         debug("%s: bitlen %d dout 0x%x din 0x%x\n",
>                 __func__, bitlen, (u32)dout, (u32)din);
> @@ -272,9 +278,17 @@ int spi_xchg_single(struct spi_slave *slave, unsigned int bitlen,
>         reg_write(&regs->ctrl, mxcs->ctrl_reg |
>                 MXC_CSPICTRL_EN | MXC_CSPICTRL_XCH);
>
> +       ts = get_timer(0);
> +       status = reg_read(&regs->stat);
>         /* Wait until the TC (Transfer completed) bit is set */
> -       while ((reg_read(&regs->stat) & MXC_CSPICTRL_TC) == 0)
> -               ;
> +       while ((status & MXC_CSPICTRL_TC) == 0) {
> +               if (get_timer(ts) > CONFIG_SYS_SPI_MXC_WAIT) {
> +                       printf("spi_xchg_single: Timeout!\n");
> +                       return -1;
> +               }
> +               udelay(10);

Why you still used explicit delay here, get_timer will do the job
finite delay right?


> +               status = reg_read(&regs->stat);
> +       }
>
>         /* Transfer completed, clear any pending request */
>         reg_write(&regs->stat, MXC_CSPICTRL_TC | MXC_CSPICTRL_RXOVF);
> --
> 1.8.3.1
>

thanks!
-- 
Jagan.