* Re: [PATCH-v2.6.39] spi/omap_mcspi: Fix broken last word xfer
2011-03-21 14:27 [PATCH-v2.6.39] spi/omap_mcspi: Fix broken last word xfer Jarkko Nikula
@ 2011-03-22 6:36 ` Poddar, Sourav
2011-03-22 8:19 ` Michael Jones
2011-03-23 21:05 ` Grant Likely
2 siblings, 0 replies; 4+ messages in thread
From: Poddar, Sourav @ 2011-03-22 6:36 UTC (permalink / raw)
To: Jarkko Nikula; +Cc: spi-devel-general, linux-omap, Grant Likely, Michael Jones
On Mon, Mar 21, 2011 at 7:57 PM, Jarkko Nikula <jhnikula@gmail.com> wrote:
> Commit adef658 "spi/omap_mcspi: catch xfers of non-multiple SPI word size"
> broke the transmission of last word in cases where access is multiple of
> word size and word size is 16 or 32 bits.
>
> Fix this by replacing the test "c > (word_len>>3)" in do-while loops with
> "c >= 'pointer increment size'". This ensures that the last word is
> transmitted in above case and still allow to break the loop and prevent
> variable c underflow in cases where word size != 'pointer increment size'.
>
> Signed-off-by: Jarkko Nikula <jhnikula@gmail.com>
> Cc: Michael Jones <michael.jones@matrix-vision.de>
> ---
> I noticed the issue with wl1251 WLAN driver on Nokia N900 using today's
> vanilla head a44f99c7efdb88fa41128065c9a9445c19894e34.
> ---
> drivers/spi/omap2_mcspi.c | 6 +++---
> 1 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/spi/omap2_mcspi.c b/drivers/spi/omap2_mcspi.c
> index 3a5ed06..6f86ba0 100644
> --- a/drivers/spi/omap2_mcspi.c
> +++ b/drivers/spi/omap2_mcspi.c
> @@ -517,7 +517,7 @@ omap2_mcspi_txrx_pio(struct spi_device *spi, struct spi_transfer *xfer)
> dev_vdbg(&spi->dev, "read-%d %02x\n",
> word_len, *(rx - 1));
> }
> - } while (c > (word_len>>3));
> + } while (c);
> } else if (word_len <= 16) {
> u16 *rx;
> const u16 *tx;
> @@ -564,7 +564,7 @@ omap2_mcspi_txrx_pio(struct spi_device *spi, struct spi_transfer *xfer)
> dev_vdbg(&spi->dev, "read-%d %04x\n",
> word_len, *(rx - 1));
> }
> - } while (c > (word_len>>3));
> + } while (c >= 2);
> } else if (word_len <= 32) {
> u32 *rx;
> const u32 *tx;
> @@ -611,7 +611,7 @@ omap2_mcspi_txrx_pio(struct spi_device *spi, struct spi_transfer *xfer)
> dev_vdbg(&spi->dev, "read-%d %08x\n",
> word_len, *(rx - 1));
> }
> - } while (c > (word_len>>3));
> + } while (c >= 4);
> }
>
> /* for TX_ONLY mode, be sure all words have shifted out */
> --
> 1.7.4.1
Latest mainline when booted on Omap4 was not booting up.
It crashes showing something like this:
[ 2.581634] ks8851 spi1.0: (unregistered net_device): spi_sync() failed
[ 2.589843] ks8851 spi1.0: (unregistered net_device): spi_sync() failed
[ 2.597900] ks8851 spi1.0: (unregistered net_device): read: spi_sync() failed
[ 2.605499] ks8851 spi1.0: failed to read device ID
The above patch when applied to the mainline solves the problem and the
kernel boots fine on Omap4.
Tested-by: Sourav Poddar<sourav.poddar@ti.com>
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH-v2.6.39] spi/omap_mcspi: Fix broken last word xfer
2011-03-21 14:27 [PATCH-v2.6.39] spi/omap_mcspi: Fix broken last word xfer Jarkko Nikula
2011-03-22 6:36 ` Poddar, Sourav
@ 2011-03-22 8:19 ` Michael Jones
2011-03-23 21:05 ` Grant Likely
2 siblings, 0 replies; 4+ messages in thread
From: Michael Jones @ 2011-03-22 8:19 UTC (permalink / raw)
To: Jarkko Nikula; +Cc: spi-devel-general, linux-omap, Grant Likely
On 03/21/2011 03:27 PM, Jarkko Nikula wrote:
> Commit adef658 "spi/omap_mcspi: catch xfers of non-multiple SPI word size"
> broke the transmission of last word in cases where access is multiple of
> word size and word size is 16 or 32 bits.
>
> Fix this by replacing the test "c > (word_len>>3)" in do-while loops with
> "c >= 'pointer increment size'". This ensures that the last word is
> transmitted in above case and still allow to break the loop and prevent
> variable c underflow in cases where word size != 'pointer increment size'.
>
> Signed-off-by: Jarkko Nikula <jhnikula@gmail.com>
> Cc: Michael Jones <michael.jones@matrix-vision.de>
> ---
> I noticed the issue with wl1251 WLAN driver on Nokia N900 using today's
> vanilla head a44f99c7efdb88fa41128065c9a9445c19894e34.
> ---
> drivers/spi/omap2_mcspi.c | 6 +++---
> 1 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/spi/omap2_mcspi.c b/drivers/spi/omap2_mcspi.c
> index 3a5ed06..6f86ba0 100644
> --- a/drivers/spi/omap2_mcspi.c
> +++ b/drivers/spi/omap2_mcspi.c
> @@ -517,7 +517,7 @@ omap2_mcspi_txrx_pio(struct spi_device *spi, struct spi_transfer *xfer)
> dev_vdbg(&spi->dev, "read-%d %02x\n",
> word_len, *(rx - 1));
> }
> - } while (c > (word_len>>3));
> + } while (c);
> } else if (word_len <= 16) {
> u16 *rx;
> const u16 *tx;
> @@ -564,7 +564,7 @@ omap2_mcspi_txrx_pio(struct spi_device *spi, struct spi_transfer *xfer)
> dev_vdbg(&spi->dev, "read-%d %04x\n",
> word_len, *(rx - 1));
> }
> - } while (c > (word_len>>3));
> + } while (c >= 2);
> } else if (word_len <= 32) {
> u32 *rx;
> const u32 *tx;
> @@ -611,7 +611,7 @@ omap2_mcspi_txrx_pio(struct spi_device *spi, struct spi_transfer *xfer)
> dev_vdbg(&spi->dev, "read-%d %08x\n",
> word_len, *(rx - 1));
> }
> - } while (c > (word_len>>3));
> + } while (c >= 4);
> }
>
> /* for TX_ONLY mode, be sure all words have shifted out */
Acked-by: Michael Jones <michael.jones@matrix-vision.de>
MATRIX VISION GmbH, Talstrasse 16, DE-71570 Oppenweiler
Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschaeftsfuehrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH-v2.6.39] spi/omap_mcspi: Fix broken last word xfer
2011-03-21 14:27 [PATCH-v2.6.39] spi/omap_mcspi: Fix broken last word xfer Jarkko Nikula
2011-03-22 6:36 ` Poddar, Sourav
2011-03-22 8:19 ` Michael Jones
@ 2011-03-23 21:05 ` Grant Likely
2 siblings, 0 replies; 4+ messages in thread
From: Grant Likely @ 2011-03-23 21:05 UTC (permalink / raw)
To: Jarkko Nikula; +Cc: spi-devel-general, linux-omap, Michael Jones
On Mon, Mar 21, 2011 at 04:27:30PM +0200, Jarkko Nikula wrote:
> Commit adef658 "spi/omap_mcspi: catch xfers of non-multiple SPI word size"
> broke the transmission of last word in cases where access is multiple of
> word size and word size is 16 or 32 bits.
>
> Fix this by replacing the test "c > (word_len>>3)" in do-while loops with
> "c >= 'pointer increment size'". This ensures that the last word is
> transmitted in above case and still allow to break the loop and prevent
> variable c underflow in cases where word size != 'pointer increment size'.
>
> Signed-off-by: Jarkko Nikula <jhnikula@gmail.com>
> Cc: Michael Jones <michael.jones@matrix-vision.de>
Applied, thanks.
g.
> ---
> I noticed the issue with wl1251 WLAN driver on Nokia N900 using today's
> vanilla head a44f99c7efdb88fa41128065c9a9445c19894e34.
> ---
> drivers/spi/omap2_mcspi.c | 6 +++---
> 1 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/spi/omap2_mcspi.c b/drivers/spi/omap2_mcspi.c
> index 3a5ed06..6f86ba0 100644
> --- a/drivers/spi/omap2_mcspi.c
> +++ b/drivers/spi/omap2_mcspi.c
> @@ -517,7 +517,7 @@ omap2_mcspi_txrx_pio(struct spi_device *spi, struct spi_transfer *xfer)
> dev_vdbg(&spi->dev, "read-%d %02x\n",
> word_len, *(rx - 1));
> }
> - } while (c > (word_len>>3));
> + } while (c);
> } else if (word_len <= 16) {
> u16 *rx;
> const u16 *tx;
> @@ -564,7 +564,7 @@ omap2_mcspi_txrx_pio(struct spi_device *spi, struct spi_transfer *xfer)
> dev_vdbg(&spi->dev, "read-%d %04x\n",
> word_len, *(rx - 1));
> }
> - } while (c > (word_len>>3));
> + } while (c >= 2);
> } else if (word_len <= 32) {
> u32 *rx;
> const u32 *tx;
> @@ -611,7 +611,7 @@ omap2_mcspi_txrx_pio(struct spi_device *spi, struct spi_transfer *xfer)
> dev_vdbg(&spi->dev, "read-%d %08x\n",
> word_len, *(rx - 1));
> }
> - } while (c > (word_len>>3));
> + } while (c >= 4);
> }
>
> /* for TX_ONLY mode, be sure all words have shifted out */
> --
> 1.7.4.1
>
^ permalink raw reply [flat|nested] 4+ messages in thread