Skip to content

C
clone-Linux

gsplhtlxg / clone-Linux

通知 2
Star 0
Fork 1
C
clone-Linux

体验新版 GitCode，发现更多精彩内容 >>

切换分支/标签
  1. 03 8月, 2010 1 次提交
  3. 08 7月, 2010 1 次提交
  5. 12 6月, 2010 1 次提交
    • K
      kbuild: Fix modpost segfault · 1c938663
      由 Krzysztof Halasa 提交于 
      Alan <alan@clueserver.org> writes:

> program: /home/alan/GitTrees/linux-2.6-mid-ref/scripts/mod/modpost -o
> Module.symvers -S vmlinux.o
>
> Program received signal SIGSEGV, Segmentation fault.

It just hit me.
It's the offset calculation in reloc_location() which overflows:
        return (void *)elf->hdr + sechdrs[section].sh_offset +
               (r->r_offset - sechdrs[section].sh_addr);

E.g. for the first rodata r entry:
r->r_offset < sechdrs[section].sh_addr
and the expression in the parenthesis produces 0xFFFFFFE0 or something
equally wise.
Reported-by: NAlan <alan@clueserver.org>
Signed-off-by: NKrzysztof Hałasa <khc@pm.waw.pl>
Tested-by: NAlan <alan@clueserver.org>
Signed-off-by: NMichal Marek <mmarek@suse.cz>
      1c938663
  7. 19 5月, 2010 1 次提交
    • O
      MODULE_DEVICE_TABLE(isapnp, ...) does nothing · fedb3d27
      由 Ondrej Zary 提交于 
      On Monday 23 November 2009 04:29:53 Rusty Russell wrote:
> On Mon, 23 Nov 2009 07:31:57 am Ondrej Zary wrote:
> > The problem is that
> > scripts/mod/file2alias.c simply ignores isapnp.
>
> AFAICT it always has, and noone has complained until now.  Perhaps
> something was still reading /lib/modules/`uname -r`/modules.isapnpmap?

The patch below works fine (at least with Debian). It needs your first
patch that moves the definitions to mod_devicetable.h. Verified that
aliases for these modules are generated correctly:

drivers/media/radio/radio-sf16fmi.c
drivers/net/ne.c
drivers/net/3c515.c
drivers/net/smc-ultra.c
drivers/pcmcia/i82365.c
drivers/scsi/aha1542.c
drivers/scsi/aha152x.c
drivers/scsi/sym53c416.c
drivers/scsi/g_NCR5380.c

Tested with RTL8019AS (ne), AVA-1505AE (aha152x) and dtc436e (g_NCR5380)
cards - they now work automatically.

Generate pnp:d aliases for isapnp_device_tables. This allows udev to load
these modules automatically.
Signed-off-by: NOndrej Zary <linux@rainbow-software.org>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
      fedb3d27
  9. 18 5月, 2010 1 次提交
  11. 03 4月, 2010 1 次提交
  13. 31 1月, 2010 4 次提交
  15. 30 1月, 2010 3 次提交
  17. 18 1月, 2010 1 次提交
    • M
      modpost: fix segfault in sym_is() with prefixed arches · 3a5dd791
      由 Mike Frysinger 提交于 
      The sym_is() compares a symbol in an attempt to automatically skip symbol
prefixes.  It does this first by searching the real symbol with the normal
unprefixed symbol.  But then it uses the length of the original symbol to
check the end of the substring instead of the length of the symbol it is
looking for.  On non-prefixed arches, this is effectively the same thing,
so there is no problem.  On prefixed-arches, since this is exceeds by just
one byte, a crash is rare and it is usually a NUL byte anyways.  But every
once in a blue moon, you get the right page alignment and it segfaults.

For example, on the Blackfin arch, sym_is() will be called with the real
symbol "___mod_usb_device_table" as "symbol" when looking for the normal
symbol "__mod_usb_device_table" as "name".  The substring will thus return
one byte into "symbol" and store it into "match".  But then "match" will
be indexed with the length of "symbol" instead of "name" and so we will
exceed the storage.  i.e. the code ends up doing:
	char foo[] = "abc"; return foo[strlen(foo)+1] == '\0';
Signed-off-by: NMike Frysinger <vapier@gentoo.org>
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      3a5dd791
  19. 15 12月, 2009 3 次提交
  21. 12 12月, 2009 2 次提交
    • N
      USB: handle bcd incrementation in usb modalias generation · 55f49f26
      由 Nathaniel McCallum 提交于 
      This patch fixes a bug when incrementing/decrementing on a BCD formatted
integer (i.e. 0x09++ should be 0x10 not 0x0A).  It just adds a function
for incrementing/decrementing BCD integers by converting to decimal,
doing the increment/decrement and then converting back to BCD.
Signed-off-by: NNathaniel McCallum <nathaniel@natemccallum.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
      55f49f26
    • N
      USB: add hex/bcd detection to usb modalias generation · afe2dab4
      由 Nathaniel McCallum 提交于 
      The current code to generate usb modaliases from usb_device_id assumes
that the device's bcdDevice descriptor will actually be in BCD format.
While this should be a sane assumption, some devices don't follow spec
and just use plain old hex.  This causes drivers for these devices to
generate invalid modalias lines which will never actually match for the
hardware.

The following patch adds hex support for bcdDevice in file2alias.c by
detecting when a driver uses a hex formatted bcdDevice_(lo|hi) and
adjusts the output to hex format accordingly.

Drivers for devices which have bcdDevice conforming to BCD will have no
change in modalias output.  Drivers for devices which don't conform
(i.e. ibmcam) should now generate valid modaliases.

EXAMPLE OUTPUT (ibmcam; space added to highlight change)
    Old: usb:v0545p800D d030[10-9] dc*dsc*dp*ic*isc*ip*
    New: usb:v0545p800D d030a      dc*dsc*dp*ic*isc*ip*
Signed-off-by: NNathaniel McCallum <nathaniel@natemccallum.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
      afe2dab4
  23. 23 9月, 2009 3 次提交
  25. 21 9月, 2009 1 次提交
  27. 12 6月, 2009 1 次提交
  29. 10 6月, 2009 2 次提交
  31. 04 5月, 2009 3 次提交
  33. 01 5月, 2009 3 次提交
    • A
      kbuild, modpost: Check the section flags, to catch missing "ax"/"aw" · b614a697
      由 Anders Kaseorg 提交于 
      When you put
  .section ".foo"
in an assembly file instead of
  .section "foo", "ax"
, one of the possible symptoms is that modpost will see an
ld-generated section name ".foo.1" in section_rel() or section_rela().
But this heuristic has two problems: it will miss a bad section that
has no relocations, and it will incorrectly flag many gcc-generated
sections as bad when compiling with -ffunction-sections
-fdata-sections.

On mips it fixes a lot of bogus warnings with gcc 4.4.0 lije this one:
WARNING: crypto/cryptd.o (.text.T.349): unexpected section name.

So instead of checking whether the section name matches a particular
pattern, we directly check for a missing SHF_ALLOC in the section
flags.
Signed-off-by: NAnders Kaseorg <andersk@mit.edu>
Tested-by: NRalf Baechle <ralf@linux-mips.org>
Signed-off-by: NSam Ravnborg <sam@ravnborg.org>
      b614a697
    • S
      kbuild: fix comment in modpost.c · c993971f
      由 Sam Ravnborg 提交于 
      There is some confusion on naming of the head section.
Correct naming is .head.text.

Fix comment so we use correct naming.
Signed-off-by: NSam Ravnborg <sam@ravnborg.org>
      c993971f
    • C
      kbuild: fix Module.markers permission error under cygwin · 99e3a1eb
      由 Cedric Hombourger 提交于 
      While building the kernel, we end-up calling modpost with -K and -M
options for the same file (Modules.markers).  This is resulting in
modpost's main function calling read_markers() and then write_markers() on
the same file.

We then have read_markers() mmap'ing the file, and writer_markers()
opening that same file for writing.

The issue is that read_markers() exits without munmap'ing the file and is
as a matter holding a reference on Modules.markers.  When write_markers()
is opening that very same file for writing, we still have a reference on
it and cygwin (Windows?) is then making fopen() fail with EPERM.

Calling release_file() before exiting read_markers() clears that reference
(and memory leak) and fopen() then succeeds.

Tested on both cygwin (1.3.22) and Linux.  Also ran modpost within
valgrind on Linux to make sure that the munmap'ed file was not accessed
after read_markers()
Signed-off-by: NCedric Hombourger <chombourger@gmail.com>
Cc: <stable@kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NSam Ravnborg <sam@ravnborg.org>
      99e3a1eb
  35. 28 4月, 2009 1 次提交
  37. 11 4月, 2009 1 次提交
  39. 31 3月, 2009 1 次提交
    • R
      module: include other structures in module version check · 8c8ef42a
      由 Rusty Russell 提交于 
      With CONFIG_MODVERSIONS, we version 'struct module' using a dummy
export, but other things matter too:

1) 'struct modversion_info' determines the layout of the __versions section,
2) 'struct kernel_param' determines the layout of the __params section,
3) 'struct kernel_symbol' determines __ksymtab*.
4) 'struct marker' determines __markers.
5) 'struct tracepoint' determines __tracepoints.

So we rename 'struct_module' to 'module_layout' and include these in
the signature.  Now it's general we can add others later on without
confusion.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
      8c8ef42a
  41. 25 3月, 2009 1 次提交
    • E
      platform: introduce module id table for platform devices · 57fee4a5
      由 Eric Miao 提交于 
      Now platform_device is being widely used on SoC processors where the
peripherals are attached to the system bus, which is simple enough.

However, silicon IPs for these SoCs are usually shared heavily across
a family of processors, even products from different companies.  This
makes the original simple driver name based matching insufficient, or
simply not straight-forward.

Introduce a module id table for platform devices, and makes it clear
that a platform driver is able to support some shared IP and handle
slight differences across different platforms (by 'driver_data').
Module alias is handled automatically when a MODULE_DEVICE_TABLE()
is defined.

To not disturb the current platform drivers too much, the matched id
entry is recorded and can be retrieved by platform_get_device_id().
Signed-off-by: NEric Miao <eric.miao@marvell.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
      57fee4a5
  43. 17 2月, 2009 1 次提交
  45. 06 2月, 2009 1 次提交
    • T
      modpost: NOBITS sections may point beyond the end of the file · 56fc82c5
      由 Tejun Heo 提交于 
      Impact: fix link failure on certain toolchains with specific configs

Recent percpu change made x86_64 split .data.init section into three
separate segments - data.init, percpu and data.init2.  data.init2 gets
.data.nosave and .bss.* and is followed by .notes segment.  Depending
on configuration both segments might contain no data, in which case
the tool chain makes the section header to contain offset beyond the
end of the file.

modpost isn't too happy about it and fails build - as reported by
Pawel Dziekonski:

    Building modules, stage 2.
    MODPOST 416 modules
    FATAL: vmlinux is truncated. sechdrs[i].sh_offset=10354688 >
    sizeof(*hrd)=64
    make[1]: *** [__modpost] Error 1

Teach modpost that NOBITS section may point beyond the end of the file
and that .modinfo can't be NOBITS.
Reported-by: NPawel Dziekonski <dzieko@gmail.com>
Signed-off-by: NTejun Heo <tj@kernel.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
      56fc82c5
  47. 21 1月, 2009 1 次提交
  49. 30 10月, 2008 1 次提交