| Message ID | 1470385072-1313-1-git-send-email-geert+renesas@glider.be (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | c3ccf357c3d75bd2924e049b6a991f7c0c111068 |
| Headers | show |
diff --git a/drivers/spi/spi-sh-msiof.c b/drivers/spi/spi-sh-msiof.c index 8d1120931f343776..91dc730880264fe8 100644 --- a/drivers/spi/spi-sh-msiof.c +++ b/drivers/spi/spi-sh-msiof.c @@ -301,6 +301,9 @@ static void sh_msiof_spi_set_clk_regs(struct sh_msiof_spi_priv *p, for (k = 0; k < ARRAY_SIZE(sh_msiof_spi_div_table); k++) { brps = DIV_ROUND_UP(div, sh_msiof_spi_div_table[k].div); + /* SCR_BRDV_DIV_1 is valid only if BRPS is x 1/1 or x 1/2 */ + if (sh_msiof_spi_div_table[k].div == 1 && brps > 2) + continue; if (brps <= 32) /* max of brdv is 32 */ break; }
The conversion from a look-up table to a calculation for clock generator parameters forgot to take into account that BRDV x 1/1 is valid only if BRPS is x 1/1 or x 1/2, leading to undefined behavior (e.g. arbitrary clock rates). This limitation is documented for the MSIOF module in all supported SH/R-Mobile and R-Car Gen2/Gen3 ARM SoCs. Tested on r8a7791/koelsch and r8a7795/salvator-x. Fixes: 65d5665bb260b034 ("spi: sh-msiof: Update calculation of frequency dividing") Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> --- I think this is suitable for stable, so please add Cc: stable@vger.kernel.org when applying. drivers/spi/spi-sh-msiof.c | 3 +++ 1 file changed, 3 insertions(+)