| Message ID | 8452191.39CeLIHTBT@wuerfel (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Fri, Dec 2, 2016 at 1:40 PM, Arnd Bergmann <arnd@arndb.de> wrote: > With binutils-2.16 and before, a weak missing symbol was kept during the 2.26? > final link, and a missing CRC for an export would lead to that CRC > being treated as zero implicitly. With binutils-2.17, the crc 2.27? > symbol gets dropped, and any module trying to use it will fail to > load. Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds -- To unsubscribe from this list: send the line "unsubscribe linux-kbuild" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Friday, December 2, 2016 1:59:15 PM CET Geert Uytterhoeven wrote: > On Fri, Dec 2, 2016 at 1:40 PM, Arnd Bergmann <arnd@arndb.de> wrote: > > With binutils-2.16 and before, a weak missing symbol was kept during the > > 2.26? > > > final link, and a missing CRC for an export would lead to that CRC > > being treated as zero implicitly. With binutils-2.17, the crc > > 2.27?' Yes, serious version number deficiency on my end. Also 4.9 instead of 3.9 in the subject... Arnd -- To unsubscribe from this list: send the line "unsubscribe linux-kbuild" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Dec 02, 2016 at 01:40:27PM +0100, Arnd Bergmann wrote: > With binutils-2.16 and before, a weak missing symbol was kept during the > final link, and a missing CRC for an export would lead to that CRC > being treated as zero implicitly. With binutils-2.17, the crc > symbol gets dropped, and any module trying to use it will fail to > load. > > This sets the weak CRC symbol to zero explicitly, making it defined > in vmlinux, which in turn lets us load the modules referring to > that CRC. > > The comment above the __CRC_SYMBOL macro suggests that this was > always the intention, although it also seems that all symbols > defined in C have a correct CRC these days, and only the exports > that are now done in assembly need this. > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > --- Looks good, works for me, and, unlike faaae2a5, doesn't produce a nasty user-scaring warning in normal operation. > diff --git a/include/asm-generic/export.h b/include/asm-generic/export.h > index 63554e9..59a3b2f 100644 > --- a/include/asm-generic/export.h > +++ b/include/asm-generic/export.h > @@ -54,6 +54,7 @@ KSYM(__kstrtab_\name): > KSYM(__kcrctab_\name): > __put KSYM(__crc_\name) > .weak KSYM(__crc_\name) > + .set KSYM(__crc_\name), 0 > .previous > #endif > #endif
On Fri, 2016-12-02 at 13:40 +0100, Arnd Bergmann wrote: > With binutils-2.16 and before, a weak missing symbol was kept during the > final link, and a missing CRC for an export would lead to that CRC > being treated as zero implicitly. With binutils-2.17, the crc > symbol gets dropped, and any module trying to use it will fail to > load. > > This sets the weak CRC symbol to zero explicitly, making it defined > in vmlinux, which in turn lets us load the modules referring to > that CRC. > > The comment above the __CRC_SYMBOL macro suggests that this was > always the intention, although it also seems that all symbols > defined in C have a correct CRC these days, and only the exports > that are now done in assembly need this. > > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > --- > Not sure if this is the correct way of doing it, but this seems trivial > enough and lets me build the kernel with missing CRCs with any binutils > version. I tried this along with Adam's patch on x86_64, with Debian's binutils 2.27.51.20161127. The result was that the kernel's __kcrctab held 0 for several symbols, even though there was type information in asm- prototypes.h and Module.symvers and the modules had a non-zero CRC for those symbols. With just Adam's patch, the kernel and modules agreed. Ben. > diff --git a/include/asm-generic/export.h b/include/asm-generic/export.h > index 63554e9..59a3b2f 100644 > --- a/include/asm-generic/export.h > +++ b/include/asm-generic/export.h > @@ -54,6 +54,7 @@ KSYM(__kstrtab_\name): > KSYM(__kcrctab_\name): > > __put KSYM(__crc_\name) > > .weak KSYM(__crc_\name) > > + .set KSYM(__crc_\name), 0 > > .previous > #endif > #endif >
On Saturday, December 3, 2016 4:36:37 AM CET Ben Hutchings wrote: > On Fri, 2016-12-02 at 13:40 +0100, Arnd Bergmann wrote: > > With binutils-2.16 and before, a weak missing symbol was kept during the > > final link, and a missing CRC for an export would lead to that CRC > > being treated as zero implicitly. With binutils-2.17, the crc > > symbol gets dropped, and any module trying to use it will fail to > > load. > > > > This sets the weak CRC symbol to zero explicitly, making it defined > > in vmlinux, which in turn lets us load the modules referring to > > that CRC. > > > > The comment above the __CRC_SYMBOL macro suggests that this was > > always the intention, although it also seems that all symbols > > defined in C have a correct CRC these days, and only the exports > > that are now done in assembly need this. > > > > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > > --- > > Not sure if this is the correct way of doing it, but this seems trivial > > enough and lets me build the kernel with missing CRCs with any binutils > > version. > > I tried this along with Adam's patch on x86_64, with Debian's binutils > 2.27.51.20161127. The result was that the kernel's __kcrctab held 0 > for several symbols, even though there was type information in asm- > prototypes.h and Module.symvers and the modules had a non-zero CRC for > those symbols. With just Adam's patch, the kernel and modules agreed. Can you be more specific? Which symbols are those? I would have expected modpost to generate Module.symvers from the vmlinux file, so I wonder where that difference comes from. Arnd -- To unsubscribe from this list: send the line "unsubscribe linux-kbuild" 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/include/asm-generic/export.h b/include/asm-generic/export.h index 63554e9..59a3b2f 100644 --- a/include/asm-generic/export.h +++ b/include/asm-generic/export.h @@ -54,6 +54,7 @@ KSYM(__kstrtab_\name): KSYM(__kcrctab_\name): __put KSYM(__crc_\name) .weak KSYM(__crc_\name) + .set KSYM(__crc_\name), 0 .previous #endif #endif
With binutils-2.16 and before, a weak missing symbol was kept during the final link, and a missing CRC for an export would lead to that CRC being treated as zero implicitly. With binutils-2.17, the crc symbol gets dropped, and any module trying to use it will fail to load. This sets the weak CRC symbol to zero explicitly, making it defined in vmlinux, which in turn lets us load the modules referring to that CRC. The comment above the __CRC_SYMBOL macro suggests that this was always the intention, although it also seems that all symbols defined in C have a correct CRC these days, and only the exports that are now done in assembly need this. Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- Not sure if this is the correct way of doing it, but this seems trivial enough and lets me build the kernel with missing CRCs with any binutils version. -- To unsubscribe from this list: send the line "unsubscribe linux-kbuild" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html