| Message ID | 20170721220142.3400093-1-arnd@arndb.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Arnd, Cyrille, I am working on fixing spi-bcm-qspi.c as per Cyrill's suggestion as mentioned here : https://patchwork.kernel.org/patch/9624585/. And remove the use of SPINOR_OP_READ* and there by remove need to include spi-nor.h. Thanks Kamal On Fri, Jul 21, 2017 at 6:00 PM, Arnd Bergmann <arnd@arndb.de> wrote: > The newly broadcom qspi driver in drivers/spi produces a build > warning when CONFIG_MTD is disabled: > > include/linux/mtd/cfi.h:76:2: #warning No CONFIG_MTD_CFI_Ix selected. No NOR chip support can work. [-Werror=cpp] > > I had suggested a workaround earlier, but Cyrille Pitchen explained > that actually the broadcom qspi is broken here and needs to be > fixed, see the lenghthy reply in patchwork. > > As nothing has happened on that driver, this tries to at least > avoid the build failure, by marking the driver as broken unless > CONFIG_MTD is enabled. Also, I add a WARN_ON_ONCE runtime > that triggers when the spi-nor framework and the driver disagree > about the command opcode, which was one of several issues that > Cyrille pointed out. > > My patch does not attempt to fix any of the actual bugs though, > it just tries to avoid the build error while highlighting the > problems. Ideally, someone would step up to create a tested > fix. If that doesn't happen, please merge my version instead > as a workaround. > > Fixes: fa236a7ef240 ("spi: bcm-qspi: Add Broadcom MSPI driver") > Link: https://patchwork.kernel.org/patch/9334097/ > Link: https://patchwork.kernel.org/patch/9624585/ > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > Cc: Cyrille Pitchen <cyrille.pitchen@atmel.com> > --- > --- > drivers/spi/Kconfig | 1 + > drivers/spi/spi-bcm-qspi.c | 23 ++++++++++++----------- > 2 files changed, 13 insertions(+), 11 deletions(-) > > diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig > index 9b31351fe429..c7a80f9d6dd0 100644 > --- a/drivers/spi/Kconfig > +++ b/drivers/spi/Kconfig > @@ -164,6 +164,7 @@ config SPI_BCM_QSPI > tristate "Broadcom BSPI and MSPI controller support" > depends on ARCH_BRCMSTB || ARCH_BCM || ARCH_BCM_IPROC || \ > BMIPS_GENERIC || COMPILE_TEST > + depends on BROKEN || MTD > default ARCH_BCM_IPROC > help > Enables support for the Broadcom SPI flash and MSPI controller. > diff --git a/drivers/spi/spi-bcm-qspi.c b/drivers/spi/spi-bcm-qspi.c > index b19722ba908c..a388a3873552 100644 > --- a/drivers/spi/spi-bcm-qspi.c > +++ b/drivers/spi/spi-bcm-qspi.c > @@ -349,12 +349,13 @@ static void bcm_qspi_bspi_set_xfer_params(struct bcm_qspi *qspi, u8 cmd_byte, > bcm_qspi_write(qspi, BSPI, BSPI_FLEX_MODE_ENABLE, flex_mode); > } > > -static int bcm_qspi_bspi_set_flex_mode(struct bcm_qspi *qspi, int width, > - int addrlen, int hp) > +static int bcm_qspi_bspi_set_flex_mode(struct bcm_qspi *qspi, > + struct spi_flash_read_message *msg, > + int width, int addrlen, int hp) > { > int bpc = 0, bpp = 0; > u8 command = SPINOR_OP_READ_FAST; > - int flex_mode = 1, rv = 0; > + int flex_mode = 1; > bool spans_4byte = false; > > dev_dbg(&qspi->pdev->dev, "set flex mode w %x addrlen %x hp %d\n", > @@ -405,15 +406,14 @@ static int bcm_qspi_bspi_set_flex_mode(struct bcm_qspi *qspi, int width, > } > break; > default: > - rv = -EINVAL; > - break; > + return -EINVAL; > } > > - if (rv == 0) > - bcm_qspi_bspi_set_xfer_params(qspi, command, bpp, bpc, > - flex_mode); > + WARN_ON_ONCE(command != msg->read_opcode); > > - return rv; > + bcm_qspi_bspi_set_xfer_params(qspi, command, bpp, bpc, > + flex_mode); > + return 0; > } > > static int bcm_qspi_bspi_set_override(struct bcm_qspi *qspi, int width, > @@ -461,6 +461,7 @@ static int bcm_qspi_bspi_set_override(struct bcm_qspi *qspi, int width, > } > > static int bcm_qspi_bspi_set_mode(struct bcm_qspi *qspi, > + struct spi_flash_read_message *msg, > int width, int addrlen, int hp) > { > int error = 0; > @@ -491,7 +492,7 @@ static int bcm_qspi_bspi_set_mode(struct bcm_qspi *qspi, > } > > if (qspi->xfer_mode.flex_mode) > - error = bcm_qspi_bspi_set_flex_mode(qspi, width, addrlen, hp); > + error = bcm_qspi_bspi_set_flex_mode(qspi, msg, width, addrlen, hp); > > if (error) { > dev_warn(&qspi->pdev->dev, > @@ -1012,7 +1013,7 @@ static int bcm_qspi_flash_read(struct spi_device *spi, > > io_width = msg->data_nbits ? msg->data_nbits : SPI_NBITS_SINGLE; > addrlen = msg->addr_width; > - ret = bcm_qspi_bspi_set_mode(qspi, io_width, addrlen, -1); > + ret = bcm_qspi_bspi_set_mode(qspi, msg, io_width, addrlen, -1); > > if (!ret) > ret = bcm_qspi_bspi_flash_read(spi, msg); > -- > 2.9.0 > -- To unsubscribe from this list: send the line "unsubscribe linux-spi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, Jul 25, 2017 at 11:39 PM, Kamal Dasu <kdasu.kdev@gmail.com> wrote: > Arnd, Cyrille, > > I am working on fixing spi-bcm-qspi.c as per Cyrill's suggestion as > mentioned here : https://patchwork.kernel.org/patch/9624585/. > And remove the use of SPINOR_OP_READ* and there by remove need to > include spi-nor.h. Ok, thanks a lot! Do you think your fix will be applicable to stable backports, or should we add a simple dependency first for the stable kernels and remove that dependency again with the proper fix? Arnd -- To unsubscribe from this list: send the line "unsubscribe linux-spi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, Jul 25, 2017 at 05:39:57PM -0400, Kamal Dasu wrote: > Arnd, Cyrille, > > I am working on fixing spi-bcm-qspi.c as per Cyrill's suggestion as > mentioned here : https://patchwork.kernel.org/patch/9624585/. > And remove the use of SPINOR_OP_READ* and there by remove need to > include spi-nor.h. That's great - any ETA? I'm tempted to apply Arnd's patch for now to help anyone who runs into problems with it, your fix could then revert it when you've got a proper fix. Please don't top post, reply in line with needed context. This allows readers to readily follow the flow of conversation and understand what you are talking about and also helps ensure that everything in the discussion is being addressed.
On Wed, Jul 26, 2017 at 3:03 AM, Arnd Bergmann <arnd@arndb.de> wrote: > On Tue, Jul 25, 2017 at 11:39 PM, Kamal Dasu <kdasu.kdev@gmail.com> wrote: >> Arnd, Cyrille, >> >> I am working on fixing spi-bcm-qspi.c as per Cyrill's suggestion as >> mentioned here : https://patchwork.kernel.org/patch/9624585/. >> And remove the use of SPINOR_OP_READ* and there by remove need to >> include spi-nor.h. > > Ok, thanks a lot! > > Do you think your fix will be applicable to stable backports, or should we > add a simple dependency first for the stable kernels and remove that > dependency again with the proper fix? > The fix would be applicable to stable backport, we don't need to add a dependency. > Arnd Kamal -- To unsubscribe from this list: send the line "unsubscribe linux-spi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig index 9b31351fe429..c7a80f9d6dd0 100644 --- a/drivers/spi/Kconfig +++ b/drivers/spi/Kconfig @@ -164,6 +164,7 @@ config SPI_BCM_QSPI tristate "Broadcom BSPI and MSPI controller support" depends on ARCH_BRCMSTB || ARCH_BCM || ARCH_BCM_IPROC || \ BMIPS_GENERIC || COMPILE_TEST + depends on BROKEN || MTD default ARCH_BCM_IPROC help Enables support for the Broadcom SPI flash and MSPI controller. diff --git a/drivers/spi/spi-bcm-qspi.c b/drivers/spi/spi-bcm-qspi.c index b19722ba908c..a388a3873552 100644 --- a/drivers/spi/spi-bcm-qspi.c +++ b/drivers/spi/spi-bcm-qspi.c @@ -349,12 +349,13 @@ static void bcm_qspi_bspi_set_xfer_params(struct bcm_qspi *qspi, u8 cmd_byte, bcm_qspi_write(qspi, BSPI, BSPI_FLEX_MODE_ENABLE, flex_mode); } -static int bcm_qspi_bspi_set_flex_mode(struct bcm_qspi *qspi, int width, - int addrlen, int hp) +static int bcm_qspi_bspi_set_flex_mode(struct bcm_qspi *qspi, + struct spi_flash_read_message *msg, + int width, int addrlen, int hp) { int bpc = 0, bpp = 0; u8 command = SPINOR_OP_READ_FAST; - int flex_mode = 1, rv = 0; + int flex_mode = 1; bool spans_4byte = false; dev_dbg(&qspi->pdev->dev, "set flex mode w %x addrlen %x hp %d\n", @@ -405,15 +406,14 @@ static int bcm_qspi_bspi_set_flex_mode(struct bcm_qspi *qspi, int width, } break; default: - rv = -EINVAL; - break; + return -EINVAL; } - if (rv == 0) - bcm_qspi_bspi_set_xfer_params(qspi, command, bpp, bpc, - flex_mode); + WARN_ON_ONCE(command != msg->read_opcode); - return rv; + bcm_qspi_bspi_set_xfer_params(qspi, command, bpp, bpc, + flex_mode); + return 0; } static int bcm_qspi_bspi_set_override(struct bcm_qspi *qspi, int width, @@ -461,6 +461,7 @@ static int bcm_qspi_bspi_set_override(struct bcm_qspi *qspi, int width, } static int bcm_qspi_bspi_set_mode(struct bcm_qspi *qspi, + struct spi_flash_read_message *msg, int width, int addrlen, int hp) { int error = 0; @@ -491,7 +492,7 @@ static int bcm_qspi_bspi_set_mode(struct bcm_qspi *qspi, } if (qspi->xfer_mode.flex_mode) - error = bcm_qspi_bspi_set_flex_mode(qspi, width, addrlen, hp); + error = bcm_qspi_bspi_set_flex_mode(qspi, msg, width, addrlen, hp); if (error) { dev_warn(&qspi->pdev->dev, @@ -1012,7 +1013,7 @@ static int bcm_qspi_flash_read(struct spi_device *spi, io_width = msg->data_nbits ? msg->data_nbits : SPI_NBITS_SINGLE; addrlen = msg->addr_width; - ret = bcm_qspi_bspi_set_mode(qspi, io_width, addrlen, -1); + ret = bcm_qspi_bspi_set_mode(qspi, msg, io_width, addrlen, -1); if (!ret) ret = bcm_qspi_bspi_flash_read(spi, msg);
The newly broadcom qspi driver in drivers/spi produces a build warning when CONFIG_MTD is disabled: include/linux/mtd/cfi.h:76:2: #warning No CONFIG_MTD_CFI_Ix selected. No NOR chip support can work. [-Werror=cpp] I had suggested a workaround earlier, but Cyrille Pitchen explained that actually the broadcom qspi is broken here and needs to be fixed, see the lenghthy reply in patchwork. As nothing has happened on that driver, this tries to at least avoid the build failure, by marking the driver as broken unless CONFIG_MTD is enabled. Also, I add a WARN_ON_ONCE runtime that triggers when the spi-nor framework and the driver disagree about the command opcode, which was one of several issues that Cyrille pointed out. My patch does not attempt to fix any of the actual bugs though, it just tries to avoid the build error while highlighting the problems. Ideally, someone would step up to create a tested fix. If that doesn't happen, please merge my version instead as a workaround. Fixes: fa236a7ef240 ("spi: bcm-qspi: Add Broadcom MSPI driver") Link: https://patchwork.kernel.org/patch/9334097/ Link: https://patchwork.kernel.org/patch/9624585/ Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Cyrille Pitchen <cyrille.pitchen@atmel.com> --- --- drivers/spi/Kconfig | 1 + drivers/spi/spi-bcm-qspi.c | 23 ++++++++++++----------- 2 files changed, 13 insertions(+), 11 deletions(-)