scripts/kallsyms: fix wrong kallsyms_relative_base with CONFIG_KALLSYMS_BASE_RELATIVE
PATCH 1/1 scripts/kallsyms: fix wrong kallsyms_relative_base with CONFIG_KALLSYMS_BASE_RELATIVE
There is the code in the read_symbol function in 'scripts/kallsyms.c':
if (is_ignored_symbol(name, type)) return NULL; /* Ignore most absolute/undefined (?) symbols. */ if (strcmp(name, "_text") == 0) _text = addr;
But the is_ignored_symbol function returns true for name="_text" and type='a'. So the next condition is not executed and the _text variable is always zero.
It makes the wrong kallsyms_relative_base symbol as a result of the code:
if (base_relative) {
output_label("kallsyms_relative_base");
output_address(relative_base);
printf("\n");
}Because the output_address function uses the _text variable.
So the kallsyms_lookup function and all related functions in the kernel do not work properly. For example, the stack trace in oops:
Call Trace: [aa095e58] [809feab8] kobj_ns_ops_tbl+0x7ff09ac8/0x7ff1c1c4 (unreliable) [aa095e98] [80002b64] kobj_ns_ops_tbl+0x7f50db74/0x80000010 [aa095ef8] [809c3d24] kobj_ns_ops_tbl+0x7feced34/0x7ff1c1c4 [aa095f28] [80002ed0] kobj_ns_ops_tbl+0x7f50dee0/0x80000010 [aa095f38] [8000f238] kobj_ns_ops_tbl+0x7f51a248/0x80000010
The right stack trace:
Call Trace: [aa095e58] [809feab8] module_vdu_video_init+0x2fc/0x3bc (unreliable) [aa095e98] [80002b64] do_one_initcall+0x40/0x1f0 [aa095ef8] [809c3d24] kernel_init_freeable+0x164/0x1d8 [aa095f28] [80002ed0] kernel_init+0x14/0x124 [aa095f38] [8000f238] ret_from_kernel_thread+0x14/0x1c
---
scripts/kallsyms.c +4 -4
@@ -195,13 +195,13 @@ static struct sym_entry *read_symbol(FILE *in)return NULL ; } -if (is_ignored_symbol (name, type)) -return NULL ; - -/* Ignore most absolute/undefined (?) symbols. */ if (strcmp (name,"_text" ) ==0 ) _text = addr; +/* Ignore most absolute/undefined (?) symbols. */ +if (is_ignored_symbol (name, type)) +return NULL ; +check_symbol_range (name, addr, text_ranges,ARRAY_SIZE (text_ranges));check_symbol_range (name, addr, &percpu_range,1 );
Hi Mikhail,
Quoted 40 lines · Mikhail Petrov
There is the code in the
read_symbolfunction in 'scripts/kallsyms.c':if (is_ignored_symbol(name, type)) return NULL; /* Ignore most absolute/undefined (?) symbols. */ if (strcmp(name, "_text") == 0) _text = addr;But the
is_ignored_symbolfunction returns true for name="_text" and type='a'. So the next condition is not executed and the_textvariable is always zero.It makes the wrong
kallsyms_relative_basesymbol as a result of the code:if (base_relative) { output_label("kallsyms_relative_base"); output_address(relative_base); printf("\n"); }Because the
output_addressfunction uses the_textvariable.So the
kallsyms_lookupfunction and all related functions in the kernel do not work properly. For example, the stack trace in oops:Call Trace: [aa095e58] [809feab8] kobj_ns_ops_tbl+0x7ff09ac8/0x7ff1c1c4 (unreliable) [aa095e98] [80002b64] kobj_ns_ops_tbl+0x7f50db74/0x80000010 [aa095ef8] [809c3d24] kobj_ns_ops_tbl+0x7feced34/0x7ff1c1c4 [aa095f28] [80002ed0] kobj_ns_ops_tbl+0x7f50dee0/0x80000010 [aa095f38] [8000f238] kobj_ns_ops_tbl+0x7f51a248/0x80000010The right stack trace:
Call Trace: [aa095e58] [809feab8] module_vdu_video_init+0x2fc/0x3bc (unreliable) [aa095e98] [80002b64] do_one_initcall+0x40/0x1f0 [aa095ef8] [809c3d24] kernel_init_freeable+0x164/0x1d8 [aa095f28] [80002ed0] kernel_init+0x14/0x124 [aa095f38] [8000f238] ret_from_kernel_thread+0x14/0x1c
Quoted 3 lines
Signed-off-by Mikhail Petrov <[email protected]>---
Thanks for the patch.
Just for curiosity, on which architecrure
did you see name="_text" and type='a' case ?
Could you wrap the commit log to avoid
this checkpatch warning?
WARNING: Possible unwrapped commit description (prefer a maximum 75 chars per line)
Also, could you shorten the patch subject
to make it fit in this limit?
Thanks.
Signature · 2 lines
Best Regards Masahiro Yamada
Hi Masahiro,
Quoted 45 lines · Masahiro Yamada
Hi Mikhail,
Quoted 42 lines · Mikhail Petrov
There is the code in the
read_symbolfunction in 'scripts/kallsyms.c':if (is_ignored_symbol(name, type)) return NULL; /* Ignore most absolute/undefined (?) symbols. */ if (strcmp(name, "_text") == 0) _text = addr;But the
is_ignored_symbolfunction returns true for name="_text" and type='a'. So the next condition is not executed and the_textvariable is always zero.It makes the wrong
kallsyms_relative_basesymbol as a result of the code:if (base_relative) { output_label("kallsyms_relative_base"); output_address(relative_base); printf("\n"); }Because the
output_addressfunction uses the_textvariable.So the
kallsyms_lookupfunction and all related functions in the kernel do not work properly. For example, the stack trace in oops:Call Trace: [aa095e58] [809feab8] kobj_ns_ops_tbl+0x7ff09ac8/0x7ff1c1c4 (unreliable) [aa095e98] [80002b64] kobj_ns_ops_tbl+0x7f50db74/0x80000010 [aa095ef8] [809c3d24] kobj_ns_ops_tbl+0x7feced34/0x7ff1c1c4 [aa095f28] [80002ed0] kobj_ns_ops_tbl+0x7f50dee0/0x80000010 [aa095f38] [8000f238] kobj_ns_ops_tbl+0x7f51a248/0x80000010The right stack trace:
Call Trace: [aa095e58] [809feab8] module_vdu_video_init+0x2fc/0x3bc (unreliable) [aa095e98] [80002b64] do_one_initcall+0x40/0x1f0 [aa095ef8] [809c3d24] kernel_init_freeable+0x164/0x1d8 [aa095f28] [80002ed0] kernel_init+0x14/0x124 [aa095f38] [8000f238] ret_from_kernel_thread+0x14/0x1cSigned-off-by Mikhail Petrov <[email protected]>
Quoted 7 lines
Quoted 1 line
---
Thanks for the patch.
Just for curiosity, on which architecrure
did you see name="_text" and type='a' case ?
Actually 'a' is 'A' (my mistake). The architecture is PowerPC - core PPC476FS.
nm -n .tmp_vmlinux1 looks like:
...
w kallsyms_token_table
w mach_powermac
00000007 a LG_CACHELINE_BYTES
00000007 a LG_CACHELINE_BYTES
00000007 a LG_CACHELINE_BYTES
00000020 a reg0000007f a CACHELINE_MASK0000007f a CACHELINE_MASK0000007f a CACHELINE_MASK
00000080 a CACHELINE_BYTES
00000080 a CACHELINE_BYTES
00000080 a CACHELINE_BYTES
00000400 a dcr
80000000 T _start
80000000 A _stext
80000000 A _text
80000088 t interrupt_base800000a0 t CriticalInput
80000180 t MachineCheck
80000260 t MachineCheckA
80000360 t DataStorage
80000420 t InstructionStorage
80000500 t ExternalInput800005c0 t Alignment
80000680 t Program
80000740 t FloatingPointUnavailable
80000820 t SystemCall
80000900 t AuxillaryProcessorUnavailable
...
Quoted 7 lines
Could you wrap the commit log to avoid
this checkpatch warning?
WARNING: Possible unwrapped commit description (prefer a maximum 75 chars per line)Also, could you shorten the patch subject
to make it fit in this limit?
Sorry for that. Now I know about scripts/checkpatch.pl. I will improve and resubmit the patch soon.
Thanks.
Quoted 70 lines · Mikhail Petrov
Hi Masahiro,
Quoted 45 lines · Masahiro Yamada
Hi Mikhail,
Quoted 42 lines · Mikhail Petrov
There is the code in the
read_symbolfunction in 'scripts/kallsyms.c':if (is_ignored_symbol(name, type)) return NULL; /* Ignore most absolute/undefined (?) symbols. */ if (strcmp(name, "_text") == 0) _text = addr;But the
is_ignored_symbolfunction returns true for name="_text" and type='a'. So the next condition is not executed and the_textvariable is always zero.It makes the wrong
kallsyms_relative_basesymbol as a result of the code:if (base_relative) { output_label("kallsyms_relative_base"); output_address(relative_base); printf("\n"); }Because the
output_addressfunction uses the_textvariable.So the
kallsyms_lookupfunction and all related functions in the kernel do not work properly. For example, the stack trace in oops:Call Trace: [aa095e58] [809feab8] kobj_ns_ops_tbl+0x7ff09ac8/0x7ff1c1c4 (unreliable) [aa095e98] [80002b64] kobj_ns_ops_tbl+0x7f50db74/0x80000010 [aa095ef8] [809c3d24] kobj_ns_ops_tbl+0x7feced34/0x7ff1c1c4 [aa095f28] [80002ed0] kobj_ns_ops_tbl+0x7f50dee0/0x80000010 [aa095f38] [8000f238] kobj_ns_ops_tbl+0x7f51a248/0x80000010The right stack trace:
Call Trace: [aa095e58] [809feab8] module_vdu_video_init+0x2fc/0x3bc (unreliable) [aa095e98] [80002b64] do_one_initcall+0x40/0x1f0 [aa095ef8] [809c3d24] kernel_init_freeable+0x164/0x1d8 [aa095f28] [80002ed0] kernel_init+0x14/0x124 [aa095f38] [8000f238] ret_from_kernel_thread+0x14/0x1cSigned-off-by Mikhail Petrov <[email protected]>Quoted 7 lines
Quoted 1 line
---
Thanks for the patch.
Just for curiosity, on which architecrure
did you see name="_text" and type='a' case ?Actually 'a' is 'A' (my mistake). The architecture is PowerPC - core PPC476FS.
nm -n .tmp_vmlinux1 looks like:
...
wkallsyms_token_table
wmach_powermac
00000007 aLG_CACHELINE_BYTES
00000007 aLG_CACHELINE_BYTES
00000007 aLG_CACHELINE_BYTES
00000020 a reg0000007faCACHELINE_MASK0000007faCACHELINE_MASK
Quoted 8 lines
0000007faCACHELINE_MASK
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000400 a dcr
80000000 T_start
80000000 A_stext
80000000 A_text
Hmm, I am still not able to reproduce this.
I compiled ARCH=powerpc, but
'powerpc-linux-nm -n .tmp_vmlinux1' got this.
0000007f a CACHELINE_MASK0000007f a CACHELINE_MASK0000007f a CACHELINE_MASK
00000080 a CACHELINE_BYTES
00000080 a CACHELINE_BYTES
00000080 a CACHELINE_BYTES
00000400 a dcrc0000000 T _startc0000000 T _stextc0000000 T _textc00000b8 t interrupt_basec00000c0 t CriticalInputc00001a0 t MachineCheckc0000280 t MachineCheckA
Which defconfig did you use?
(I also CCed the ppc maintainer,
I am just curious what makes _text absolute.)
Quoted 20 lines
80000088 t
interrupt_base800000a0t CriticalInput
80000180 t MachineCheck
80000260 t MachineCheckA
80000360 t DataStorage
80000420 t InstructionStorage
80000500 t ExternalInput800005c0t Alignment
80000680 t Program
80000740 t FloatingPointUnavailable
80000820 t SystemCall
80000900 t AuxillaryProcessorUnavailable
...Quoted 5 lines
Could you wrap the commit log to avoid
this checkpatch warning?
WARNING: Possible unwrapped commit description (prefer a maximum 75 chars per line)
Quoted 6 lines
Quoted 2 lines
Also, could you shorten the patch subject
to make it fit in this limit?Sorry for that. Now I know about
scripts/checkpatch.pl. I will improve and resubmit the patch soon.Thanks.
--
Best Regards
Masahiro Yamada
Quoted 93 lines · Masahiro Yamada
Quoted 60 lines · Mikhail Petrov
Quoted 41 lines · Masahiro Yamada
Quoted 31 lines · Mikhail Petrov
There is the code in the
read_symbolfunction in 'scripts/kallsyms.c':if (is_ignored_symbol(name, type)) return NULL; /* Ignore most absolute/undefined (?) symbols. */ if (strcmp(name, "_text") == 0) _text = addr;But the
is_ignored_symbolfunction returns true for name="_text" and type='a'. So the next condition is not executed and the_textvariable is always zero.It makes the wrong
kallsyms_relative_basesymbol as a result of the code:if (base_relative) { output_label("kallsyms_relative_base"); output_address(relative_base); printf("\n"); }Because the
output_addressfunction uses the_textvariable.So the
kallsyms_lookupfunction and all related functions in the kernel do not work properly. For example, the stack trace in oops:Call Trace: [aa095e58] [809feab8] kobj_ns_ops_tbl+0x7ff09ac8/0x7ff1c1c4 (unreliable) [aa095e98] [80002b64] kobj_ns_ops_tbl+0x7f50db74/0x80000010 [aa095ef8] [809c3d24] kobj_ns_ops_tbl+0x7feced34/0x7ff1c1c4 [aa095f28] [80002ed0] kobj_ns_ops_tbl+0x7f50dee0/0x80000010 [aa095f38] [8000f238] kobj_ns_ops_tbl+0x7f51a248/0x80000010Quoted 8 lines
The right stack trace:
Call Trace: [aa095e58] [809feab8] module_vdu_video_init+0x2fc/0x3bc (unreliable) [aa095e98] [80002b64] do_one_initcall+0x40/0x1f0 [aa095ef8] [809c3d24] kernel_init_freeable+0x164/0x1d8 [aa095f28] [80002ed0] kernel_init+0x14/0x124 [aa095f38] [8000f238] ret_from_kernel_thread+0x14/0x1cQuoted 4 lines
Thanks for the patch.
Just for curiosity, on which architecrure
did you see name="_text" and type='a' case ?Actually 'a' is 'A' (my mistake). The architecture is PowerPC - core PPC476FS.
nm -n .tmp_vmlinux1 looks like:
...
wkallsyms_token_table
wmach_powermac
00000007 aLG_CACHELINE_BYTES
00000007 aLG_CACHELINE_BYTES
00000007 aLG_CACHELINE_BYTES
00000020 a reg0000007faCACHELINE_MASK0000007faCACHELINE_MASKQuoted 8 lines
0000007faCACHELINE_MASK
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000400 a dcr
80000000 T_start
80000000 A_stext
80000000 A_textHmm, I am still not able to reproduce this.
I compiled ARCH=powerpc, but
'powerpc-linux-nm -n .tmp_vmlinux1' got this.
0000007faCACHELINE_MASK0000007faCACHELINE_MASK0000007faCACHELINE_MASK
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000400 a dcrc0000000T_startc0000000T_stextc0000000T_textc00000b8tinterrupt_basec00000c0t CriticalInputc00001a0t MachineCheckc0000280t MachineCheckA
Quoted 5 lines
Which defconfig did you use?
(I also CCed the ppc maintainer,
I am just curious what makes_textabsolute.)
I have no idea sorry.
arch/powerpc has about 20 sub-platforms that do weird and wonderful things. Presumably this happens on one of those.
I played around with some of the defconfigs but couldn't reproduce this.
The only config we have that puts the kernel at 0x80000000 is:
$ git grep 80000000 arch/powerpc/configs/ arch/powerpc/configs/85xx/xes_mpc85xx_defconfig:CONFIG_PAGE_OFFSET=0x80000000
But that's not a PPC476 platform.
And _text is not 'A':
$ nm .tmp_vmlinux1 | grep -w _text 80000000 T _text
cheers
Quoted 105 lines · Masahiro Yamada
Quoted 70 lines · Mikhail Petrov
Hi Masahiro,
Quoted 45 lines · Masahiro Yamada
Hi Mikhail,
Quoted 42 lines · Mikhail Petrov
There is the code in the
read_symbolfunction in 'scripts/kallsyms.c':if (is_ignored_symbol(name, type)) return NULL; /* Ignore most absolute/undefined (?) symbols. */ if (strcmp(name, "_text") == 0) _text = addr;But the
is_ignored_symbolfunction returns true for name="_text" and type='a'. So the next condition is not executed and the_textvariable is always zero.It makes the wrong
kallsyms_relative_basesymbol as a result of the code:if (base_relative) { output_label("kallsyms_relative_base"); output_address(relative_base); printf("\n"); }Because the
output_addressfunction uses the_textvariable.So the
kallsyms_lookupfunction and all related functions in the kernel do not work properly. For example, the stack trace in oops:Call Trace: [aa095e58] [809feab8] kobj_ns_ops_tbl+0x7ff09ac8/0x7ff1c1c4 (unreliable) [aa095e98] [80002b64] kobj_ns_ops_tbl+0x7f50db74/0x80000010 [aa095ef8] [809c3d24] kobj_ns_ops_tbl+0x7feced34/0x7ff1c1c4 [aa095f28] [80002ed0] kobj_ns_ops_tbl+0x7f50dee0/0x80000010 [aa095f38] [8000f238] kobj_ns_ops_tbl+0x7f51a248/0x80000010The right stack trace:
Call Trace: [aa095e58] [809feab8] module_vdu_video_init+0x2fc/0x3bc (unreliable) [aa095e98] [80002b64] do_one_initcall+0x40/0x1f0 [aa095ef8] [809c3d24] kernel_init_freeable+0x164/0x1d8 [aa095f28] [80002ed0] kernel_init+0x14/0x124 [aa095f38] [8000f238] ret_from_kernel_thread+0x14/0x1cSigned-off-by Mikhail Petrov <[email protected]>Quoted 7 lines
Quoted 1 line
---
Thanks for the patch.
Just for curiosity, on which architecrure
did you see name="_text" and type='a' case ?Actually 'a' is 'A' (my mistake). The architecture is PowerPC - core PPC476FS.
nm -n .tmp_vmlinux1 looks like:
...
wkallsyms_token_table
wmach_powermac
00000007 aLG_CACHELINE_BYTES
00000007 aLG_CACHELINE_BYTES
00000007 aLG_CACHELINE_BYTES
00000020 a reg0000007faCACHELINE_MASK0000007faCACHELINE_MASKQuoted 8 lines
0000007faCACHELINE_MASK
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000400 a dcr
80000000 T_start
80000000 A_stext
80000000 A_textHmm, I am still not able to reproduce this.
I compiled ARCH=powerpc, but
'powerpc-linux-nm -n .tmp_vmlinux1' got this.
0000007faCACHELINE_MASK0000007faCACHELINE_MASK0000007faCACHELINE_MASK
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000400 a dcrc0000000T_startc0000000T_stextc0000000T_textc00000b8tinterrupt_basec00000c0t CriticalInputc00001a0t MachineCheckc0000280t MachineCheckA
Quoted 1 line
Which defconfig did you use?
I use a custom config file for a custom SoC with two PPC476FS cores. The config is not in the upstream repository. The same effect can be reached with '44x/akebono_defconfig'.
I did some investigation with the GCC version.
GCC version 4.8.1 (akebono_defconfig):
00000007 a LG_CACHELINE_BYTES
00000020 a reg0000007f a CACHELINE_MASK0000007f a CACHELINE_MASK0000007f a CACHELINE_MASK
00000080 a CACHELINE_BYTES
00000080 a CACHELINE_BYTES
00000080 a CACHELINE_BYTES
00000400 a dcr
00000400 a sprc0000000 T _startc0000000 A _stextc0000000 A _textc0000088 t interrupt_basec00000a0 t CriticalInputc0000180 t MachineCheck
GCC version 7.5.0 (akebono_defconfig):
00000007 a LG_CACHELINE_BYTES
00000020 a reg0000007f a CACHELINE_MASK0000007f a CACHELINE_MASK0000007f a CACHELINE_MASK
00000080 a CACHELINE_BYTES
00000080 a CACHELINE_BYTES
00000080 a CACHELINE_BYTES
00000400 a dcr
00000400 a sprc0000000 T _startc0000000 T _stextc0000000 T _textc0000088 t interrupt_basec00000a0 t CriticalInputc0000180 t MachineCheck
So, I used an old version of GCC. Changing the GCC version solved the problem. Maybe the patch is not necessary.
Quoted 34 lines
(I also CCed the ppc maintainer,
I am just curious what makes_textabsolute.)Quoted 23 lines
80000088 t
interrupt_base800000a0t CriticalInput
80000180 t MachineCheck
80000260 t MachineCheckA
80000360 t DataStorage
80000420 t InstructionStorage
80000500 t ExternalInput800005c0t Alignment
80000680 t Program
80000740 t FloatingPointUnavailable
80000820 t SystemCall
80000900 t AuxillaryProcessorUnavailable
...Quoted 7 lines
Could you wrap the commit log to avoid
this checkpatch warning?
WARNING: Possible unwrapped commit description (prefer a maximum 75 chars per line)Also, could you shorten the patch subject
to make it fit in this limit?
Quoted 9 lines
Quoted 3 lines
Sorry for that. Now I know about
scripts/checkpatch.pl. I will improve and resubmit the patch soon.Thanks.
--
Best Regards
Masahiro Yamada
--
Best Regards,
Mikhail Petrov
Hi Michael,
Quoted 105 lines · Michael Ellerman
Quoted 93 lines · Masahiro Yamada
Quoted 60 lines · Mikhail Petrov
Quoted 41 lines · Masahiro Yamada
Quoted 31 lines · Mikhail Petrov
There is the code in the
read_symbolfunction in 'scripts/kallsyms.c':if (is_ignored_symbol(name, type)) return NULL; /* Ignore most absolute/undefined (?) symbols. */ if (strcmp(name, "_text") == 0) _text = addr;But the
is_ignored_symbolfunction returns true for name="_text" and type='a'. So the next condition is not executed and the_textvariable is always zero.It makes the wrong
kallsyms_relative_basesymbol as a result of the code:if (base_relative) { output_label("kallsyms_relative_base"); output_address(relative_base); printf("\n"); }Because the
output_addressfunction uses the_textvariable.So the
kallsyms_lookupfunction and all related functions in the kernel do not work properly. For example, the stack trace in oops:Call Trace: [aa095e58] [809feab8] kobj_ns_ops_tbl+0x7ff09ac8/0x7ff1c1c4 (unreliable) [aa095e98] [80002b64] kobj_ns_ops_tbl+0x7f50db74/0x80000010 [aa095ef8] [809c3d24] kobj_ns_ops_tbl+0x7feced34/0x7ff1c1c4 [aa095f28] [80002ed0] kobj_ns_ops_tbl+0x7f50dee0/0x80000010 [aa095f38] [8000f238] kobj_ns_ops_tbl+0x7f51a248/0x80000010Quoted 8 lines
The right stack trace:
Call Trace: [aa095e58] [809feab8] module_vdu_video_init+0x2fc/0x3bc (unreliable) [aa095e98] [80002b64] do_one_initcall+0x40/0x1f0 [aa095ef8] [809c3d24] kernel_init_freeable+0x164/0x1d8 [aa095f28] [80002ed0] kernel_init+0x14/0x124 [aa095f38] [8000f238] ret_from_kernel_thread+0x14/0x1cQuoted 4 lines
Thanks for the patch.
Just for curiosity, on which architecrure
did you see name="_text" and type='a' case ?Actually 'a' is 'A' (my mistake). The architecture is PowerPC - core PPC476FS.
nm -n .tmp_vmlinux1 looks like:
...
wkallsyms_token_table
wmach_powermac
00000007 aLG_CACHELINE_BYTES
00000007 aLG_CACHELINE_BYTES
00000007 aLG_CACHELINE_BYTES
00000020 a reg0000007faCACHELINE_MASK0000007faCACHELINE_MASKQuoted 8 lines
0000007faCACHELINE_MASK
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000400 a dcr
80000000 T_start
80000000 A_stext
80000000 A_textHmm, I am still not able to reproduce this.
I compiled ARCH=powerpc, but
'powerpc-linux-nm -n .tmp_vmlinux1' got this.
0000007faCACHELINE_MASK0000007faCACHELINE_MASK0000007faCACHELINE_MASK
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000400 a dcrc0000000T_startc0000000T_stextc0000000T_textc00000b8tinterrupt_basec00000c0t CriticalInputc00001a0t MachineCheckc0000280t MachineCheckAQuoted 5 lines
Which defconfig did you use?
(I also CCed the ppc maintainer,
I am just curious what makes_textabsolute.)I have no idea sorry.
arch/powerpchas about 20 sub-platforms that do weird and wonderful things. Presumably this happens on one of those.
Quoted 8 lines
I played around with some of the defconfigs but couldn't reproduce this.
The only config we have that puts the kernel at
0x80000000is:$ git grep 80000000 arch/powerpc/configs/ arch/powerpc/configs/85xx/xes_mpc85xx_defconfig:CONFIG_PAGE_OFFSET=0x80000000But that's not a PPC476 platform.
It is a custom config for a custom SoC. The config is not in the upstream repository. PAGE_OFFSET has been changed for increasing the address space for PCI windows and some devices.
Quoted 5 lines
And
_textis not 'A':$ nm .tmp_vmlinux1 | grep -w _text 80000000 T _text
I used the old GCC version - 4.8.1. In modern versions of GCC, the record has type 'T'.
Quoted 4 lines
cheers
--
Best Regards
Mikhail Petrov
Hi Mikhail,
Quoted 153 lines · Mikhail Petrov
Quoted 105 lines · Masahiro Yamada
Quoted 70 lines · Mikhail Petrov
Hi Masahiro,
Quoted 45 lines · Masahiro Yamada
Hi Mikhail,
Quoted 42 lines · Mikhail Petrov
There is the code in the
read_symbolfunction in 'scripts/kallsyms.c':if (is_ignored_symbol(name, type)) return NULL; /* Ignore most absolute/undefined (?) symbols. */ if (strcmp(name, "_text") == 0) _text = addr;But the
is_ignored_symbolfunction returns true for name="_text" and type='a'. So the next condition is not executed and the_textvariable is always zero.It makes the wrong
kallsyms_relative_basesymbol as a result of the code:if (base_relative) { output_label("kallsyms_relative_base"); output_address(relative_base); printf("\n"); }Because the
output_addressfunction uses the_textvariable.So the
kallsyms_lookupfunction and all related functions in the kernel do not work properly. For example, the stack trace in oops:Call Trace: [aa095e58] [809feab8] kobj_ns_ops_tbl+0x7ff09ac8/0x7ff1c1c4 (unreliable) [aa095e98] [80002b64] kobj_ns_ops_tbl+0x7f50db74/0x80000010 [aa095ef8] [809c3d24] kobj_ns_ops_tbl+0x7feced34/0x7ff1c1c4 [aa095f28] [80002ed0] kobj_ns_ops_tbl+0x7f50dee0/0x80000010 [aa095f38] [8000f238] kobj_ns_ops_tbl+0x7f51a248/0x80000010The right stack trace:
Call Trace: [aa095e58] [809feab8] module_vdu_video_init+0x2fc/0x3bc (unreliable) [aa095e98] [80002b64] do_one_initcall+0x40/0x1f0 [aa095ef8] [809c3d24] kernel_init_freeable+0x164/0x1d8 [aa095f28] [80002ed0] kernel_init+0x14/0x124 [aa095f38] [8000f238] ret_from_kernel_thread+0x14/0x1cSigned-off-by Mikhail Petrov <[email protected]>Quoted 7 lines
Quoted 1 line
---
Thanks for the patch.
Just for curiosity, on which architecrure
did you see name="_text" and type='a' case ?Actually 'a' is 'A' (my mistake). The architecture is PowerPC - core PPC476FS.
nm -n .tmp_vmlinux1 looks like:
...
wkallsyms_token_table
wmach_powermac
00000007 aLG_CACHELINE_BYTES
00000007 aLG_CACHELINE_BYTES
00000007 aLG_CACHELINE_BYTES
00000020 a reg0000007faCACHELINE_MASK0000007faCACHELINE_MASKQuoted 8 lines
0000007faCACHELINE_MASK
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000400 a dcr
80000000 T_start
80000000 A_stext
80000000 A_textHmm, I am still not able to reproduce this.
I compiled ARCH=powerpc, but
'powerpc-linux-nm -n .tmp_vmlinux1' got this.
0000007faCACHELINE_MASK0000007faCACHELINE_MASK0000007faCACHELINE_MASK
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000400 a dcrc0000000T_startc0000000T_stextc0000000T_textc00000b8tinterrupt_basec00000c0t CriticalInputc00001a0t MachineCheckc0000280t MachineCheckAQuoted 1 line
Which defconfig did you use?
I use a custom config file for a custom SoC with two PPC476FS cores. The config is not in the upstream repository. The same effect can be reached with '44x/akebono_defconfig'.
I did some investigation with the GCC version.
GCC version 4.8.1 (
akebono_defconfig):00000007 a
LG_CACHELINE_BYTES
00000020 a reg0000007faCACHELINE_MASK0000007faCACHELINE_MASK0000007faCACHELINE_MASK
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000400 a dcr
00000400 a sprc0000000T_startc0000000A_stextc0000000A_textc0000088tinterrupt_basec00000a0t CriticalInputc0000180t MachineCheckGCC version 7.5.0 (
akebono_defconfig):00000007 a
LG_CACHELINE_BYTES
00000020 a reg0000007faCACHELINE_MASK0000007faCACHELINE_MASK0000007faCACHELINE_MASK
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000080 aCACHELINE_BYTES
00000400 a dcr
00000400 a sprc0000000T_startc0000000T_stextc0000000T_textc0000088tinterrupt_basec00000a0t CriticalInputc0000180t MachineCheck
Quoted 1 line
So, I used an old version of GCC. Changing the GCC version solved the problem. Maybe the patch is not necessary.
Confirmed.
I was able to reproduce it with the following toolchain.
Actually, this is not GCC, but linker (binutils) issue.
This happens on any architecture
if you use GNU binutils <= 2.22
The current minimal supported version is
2.21, so I will pick up v2.
Quoted 40 lines
Quoted 40 lines
(I also CCed the ppc maintainer,
I am just curious what makes_textabsolute.)Quoted 26 lines
80000088 t
interrupt_base800000a0t CriticalInput
80000180 t MachineCheck
80000260 t MachineCheckA
80000360 t DataStorage
80000420 t InstructionStorage
80000500 t ExternalInput800005c0t Alignment
80000680 t Program
80000740 t FloatingPointUnavailable
80000820 t SystemCall
80000900 t AuxillaryProcessorUnavailable
...Quoted 7 lines
Could you wrap the commit log to avoid
this checkpatch warning?
WARNING: Possible unwrapped commit description (prefer a maximum 75 chars per line)Also, could you shorten the patch subject
to make it fit in this limit?Sorry for that. Now I know about
scripts/checkpatch.pl. I will improve and resubmit the patch soon.Thanks.
Quoted 9 lines
Quoted 4 lines
--
Best Regards
Masahiro Yamada--
Best Regards,
Mikhail Petrov
--
Best Regards
Masahiro Yamada