Message ID | 1445437906-23633-1-git-send-email-linux@roeck-us.net |
---|---|
State | Superseded |
Headers | show |
+ Han On Wed, Oct 21, 2015 at 07:31:46AM -0700, Guenter Roeck wrote: > Attempts to build fsl-quadspi on SPARC fail with > > drivers/mtd/spi-nor/fsl-quadspi.c: In function 'fsl_qspi_init_lut': > drivers/mtd/spi-nor/fsl-quadspi.c:369:1: error: > 'LUT_0' undeclared (first use in this function) > drivers/mtd/spi-nor/fsl-quadspi.c:418:1: error: > pasting "LUT_" and "(" does not give a valid preprocessing token > drivers/mtd/spi-nor/fsl-quadspi.c:418:2: error: > implicit declaration of function 'LUT_' I don't think this is only a SPARC problem. The macro concatenation is generally suspect. I see that READ and WRITE are problems at least. If something like <linux/fs.h> gets included somehow, then these tokens resolve to integers or expressions before they fall through to literal concatentation, so we get 'LUT_0' or 'LUT_(1ULL << __REQ_WRITE)' instead of 'LUT_READ' and 'LUT_WRITE'. Han, are you up for fixing, or should I? Brian > Never build it on SPARC, even if COMPILE_TEST is enabled. > > Fixes: d26a22d06708 ("mtd: fsl-quadspi: allow building for other ARCHes > with COMPILE_TEST") > Cc: Brian Norris <computersforpeace@gmail.com> > Signed-off-by: Guenter Roeck <linux@roeck-us.net> > --- > drivers/mtd/spi-nor/Kconfig | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/mtd/spi-nor/Kconfig b/drivers/mtd/spi-nor/Kconfig > index 2fe2a7e90fa9..fceb2886bfd1 100644 > --- a/drivers/mtd/spi-nor/Kconfig > +++ b/drivers/mtd/spi-nor/Kconfig > @@ -23,7 +23,7 @@ config MTD_SPI_NOR_USE_4K_SECTORS > > config SPI_FSL_QUADSPI > tristate "Freescale Quad SPI controller" > - depends on ARCH_MXC || COMPILE_TEST > + depends on ARCH_MXC || (COMPILE_TEST && !SPARC) > depends on HAS_IOMEM > help > This enables support for the Quad SPI controller in master mode. > -- > 2.1.4 >
On 10/22/2015 06:07 PM, Brian Norris wrote: > + Han > > On Wed, Oct 21, 2015 at 07:31:46AM -0700, Guenter Roeck wrote: >> Attempts to build fsl-quadspi on SPARC fail with >> >> drivers/mtd/spi-nor/fsl-quadspi.c: In function 'fsl_qspi_init_lut': >> drivers/mtd/spi-nor/fsl-quadspi.c:369:1: error: >> 'LUT_0' undeclared (first use in this function) >> drivers/mtd/spi-nor/fsl-quadspi.c:418:1: error: >> pasting "LUT_" and "(" does not give a valid preprocessing token >> drivers/mtd/spi-nor/fsl-quadspi.c:418:2: error: >> implicit declaration of function 'LUT_' > > I don't think this is only a SPARC problem. The macro concatenation is > generally suspect. > > I see that READ and WRITE are problems at least. If something like > <linux/fs.h> gets included somehow, then these tokens resolve to > integers or expressions before they fall through to literal > concatentation, so we get 'LUT_0' or 'LUT_(1ULL << __REQ_WRITE)' instead > of 'LUT_READ' and 'LUT_WRITE'. > You are right, that deserves a better fix. I'll leave it up to you. Thanks, Guenter
On Thu, Oct 22, 2015 at 8:13 PM, Guenter Roeck <linux@roeck-us.net> wrote: > On 10/22/2015 06:07 PM, Brian Norris wrote: >> >> + Han >> >> On Wed, Oct 21, 2015 at 07:31:46AM -0700, Guenter Roeck wrote: >>> >>> Attempts to build fsl-quadspi on SPARC fail with >>> >>> drivers/mtd/spi-nor/fsl-quadspi.c: In function 'fsl_qspi_init_lut': >>> drivers/mtd/spi-nor/fsl-quadspi.c:369:1: error: >>> 'LUT_0' undeclared (first use in this function) >>> drivers/mtd/spi-nor/fsl-quadspi.c:418:1: error: >>> pasting "LUT_" and "(" does not give a valid preprocessing token >>> drivers/mtd/spi-nor/fsl-quadspi.c:418:2: error: >>> implicit declaration of function 'LUT_' >> >> >> I don't think this is only a SPARC problem. The macro concatenation is >> generally suspect. >> >> I see that READ and WRITE are problems at least. If something like >> <linux/fs.h> gets included somehow, then these tokens resolve to >> integers or expressions before they fall through to literal >> concatentation, so we get 'LUT_0' or 'LUT_(1ULL << __REQ_WRITE)' instead >> of 'LUT_READ' and 'LUT_WRITE'. Hi Brian We have two options, either undef READ and WRITE before these macros or change the LUT_READ to LUT_FSLREAD and same as WRITE. So what's your opinion? Thanks. >> > You are right, that deserves a better fix. I'll leave it up to you. > > Thanks, > Guenter > > > > ______________________________________________________ > Linux MTD discussion mailing list > http://lists.infradead.org/mailman/listinfo/linux-mtd/
On 10/23/2015 09:31 AM, Han Xu wrote: > On Thu, Oct 22, 2015 at 8:13 PM, Guenter Roeck <linux@roeck-us.net> wrote: >> On 10/22/2015 06:07 PM, Brian Norris wrote: >>> >>> + Han >>> >>> On Wed, Oct 21, 2015 at 07:31:46AM -0700, Guenter Roeck wrote: >>>> >>>> Attempts to build fsl-quadspi on SPARC fail with >>>> >>>> drivers/mtd/spi-nor/fsl-quadspi.c: In function 'fsl_qspi_init_lut': >>>> drivers/mtd/spi-nor/fsl-quadspi.c:369:1: error: >>>> 'LUT_0' undeclared (first use in this function) >>>> drivers/mtd/spi-nor/fsl-quadspi.c:418:1: error: >>>> pasting "LUT_" and "(" does not give a valid preprocessing token >>>> drivers/mtd/spi-nor/fsl-quadspi.c:418:2: error: >>>> implicit declaration of function 'LUT_' >>> >>> >>> I don't think this is only a SPARC problem. The macro concatenation is >>> generally suspect. >>> >>> I see that READ and WRITE are problems at least. If something like >>> <linux/fs.h> gets included somehow, then these tokens resolve to >>> integers or expressions before they fall through to literal >>> concatentation, so we get 'LUT_0' or 'LUT_(1ULL << __REQ_WRITE)' instead >>> of 'LUT_READ' and 'LUT_WRITE'. > Hi Brian > We have two options, either undef READ and WRITE before these macros > or change the LUT_READ to LUT_FSLREAD and same as WRITE. So what's > your opinion? Thanks. Undef sounds like a really bad solution to me. Guenter >>> >> You are right, that deserves a better fix. I'll leave it up to you. >> >> Thanks, >> Guenter >> >> >> >> ______________________________________________________ >> Linux MTD discussion mailing list >> http://lists.infradead.org/mailman/listinfo/linux-mtd/ >
On Fri, Oct 23, 2015 at 11:31:47AM -0500, Han Xu wrote: > On Thu, Oct 22, 2015 at 8:13 PM, Guenter Roeck <linux@roeck-us.net> wrote: > > On 10/22/2015 06:07 PM, Brian Norris wrote: > >> > >> + Han > >> > >> On Wed, Oct 21, 2015 at 07:31:46AM -0700, Guenter Roeck wrote: > >>> > >>> Attempts to build fsl-quadspi on SPARC fail with > >>> > >>> drivers/mtd/spi-nor/fsl-quadspi.c: In function 'fsl_qspi_init_lut': > >>> drivers/mtd/spi-nor/fsl-quadspi.c:369:1: error: > >>> 'LUT_0' undeclared (first use in this function) > >>> drivers/mtd/spi-nor/fsl-quadspi.c:418:1: error: > >>> pasting "LUT_" and "(" does not give a valid preprocessing token > >>> drivers/mtd/spi-nor/fsl-quadspi.c:418:2: error: > >>> implicit declaration of function 'LUT_' > >> > >> > >> I don't think this is only a SPARC problem. The macro concatenation is > >> generally suspect. > >> > >> I see that READ and WRITE are problems at least. If something like > >> <linux/fs.h> gets included somehow, then these tokens resolve to > >> integers or expressions before they fall through to literal > >> concatentation, so we get 'LUT_0' or 'LUT_(1ULL << __REQ_WRITE)' instead > >> of 'LUT_READ' and 'LUT_WRITE'. > Hi Brian > We have two options, either undef READ and WRITE before these macros > or change the LUT_READ to LUT_FSLREAD and same as WRITE. So what's > your opinion? Thanks. I'd rule out #undef'ing other macros. That just looks like a hack. There's at least one other option: de-obfuscate your code by directly using macros instead of concatenating LUT_ to save a few characters. So: #define LUT0(ins, pad, opr) \ (((opr) << OPRND0_SHIFT) | ((pad) << PAD0_SHIFT) | \ ((ins) << INSTR0_SHIFT)) and for example: writel(LUT0(LUT_WRITE, LUT_PAD1, 0), base + QUADSPI_LUT(lut_base + 1)); I'd take either that option, or renaming your LUT_READ and LUT_WRITE to LUT_FSL_READ and LUT_FSL_WRITE (not LUT_FSLREAD and LUT_FSLWRITE). Brian
diff --git a/drivers/mtd/spi-nor/Kconfig b/drivers/mtd/spi-nor/Kconfig index 2fe2a7e90fa9..fceb2886bfd1 100644 --- a/drivers/mtd/spi-nor/Kconfig +++ b/drivers/mtd/spi-nor/Kconfig @@ -23,7 +23,7 @@ config MTD_SPI_NOR_USE_4K_SECTORS config SPI_FSL_QUADSPI tristate "Freescale Quad SPI controller" - depends on ARCH_MXC || COMPILE_TEST + depends on ARCH_MXC || (COMPILE_TEST && !SPARC) depends on HAS_IOMEM help This enables support for the Quad SPI controller in master mode.
Attempts to build fsl-quadspi on SPARC fail with drivers/mtd/spi-nor/fsl-quadspi.c: In function 'fsl_qspi_init_lut': drivers/mtd/spi-nor/fsl-quadspi.c:369:1: error: 'LUT_0' undeclared (first use in this function) drivers/mtd/spi-nor/fsl-quadspi.c:418:1: error: pasting "LUT_" and "(" does not give a valid preprocessing token drivers/mtd/spi-nor/fsl-quadspi.c:418:2: error: implicit declaration of function 'LUT_' Never build it on SPARC, even if COMPILE_TEST is enabled. Fixes: d26a22d06708 ("mtd: fsl-quadspi: allow building for other ARCHes with COMPILE_TEST") Cc: Brian Norris <computersforpeace@gmail.com> Signed-off-by: Guenter Roeck <linux@roeck-us.net> --- drivers/mtd/spi-nor/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)