Re: [PATCH] kallsyms, tracing: output more proper symbol name

[Frederic Weisbecker]

On Wed, Mar 11, 2009 at 10:46:21AM +0100, Ingo Molnar wrote:
> 
> * Lai Jiangshan <laijs@xxxxxxxxxxxxxx> wrote:
> 
> > +#define VMLINUX_SYMBOL(_sym_) #_sym_
> > +
> > +static const char *hide_symbols[] = {
> > +	/* misc symbols */
> > +	"_text",
> > +	"_stext",
> > +	"_etext",
> > +	"_sinittext",
> > +	"_einittext",
> > +
> > +	/* symbols from include/asm-generic/vmlinux.lds.h */
> > +	VMLINUX_SYMBOL(__start_mcount_loc),
> > +	VMLINUX_SYMBOL(__stop_mcount_loc),
> > +	VMLINUX_SYMBOL(__start_annotated_branch_profile),
> > +	VMLINUX_SYMBOL(__stop_annotated_branch_profile),
> > +	VMLINUX_SYMBOL(__start_branch_profile),
> > +	VMLINUX_SYMBOL(__stop_branch_profile),
> > +	VMLINUX_SYMBOL(__start___markers),
> > +	VMLINUX_SYMBOL(__stop___markers),
> > +	VMLINUX_SYMBOL(__start___tracepoints),
> > +	VMLINUX_SYMBOL(__stop___tracepoints),
> > +	VMLINUX_SYMBOL(__start_rodata),
> > +	VMLINUX_SYMBOL(__start_pci_fixups_early),
> > +	VMLINUX_SYMBOL(__end_pci_fixups_early),
> > +	VMLINUX_SYMBOL(__start_pci_fixups_header),
> > +	VMLINUX_SYMBOL(__end_pci_fixups_header),
> > +	VMLINUX_SYMBOL(__start_pci_fixups_final),
> > +	VMLINUX_SYMBOL(__end_pci_fixups_final),
> > +	VMLINUX_SYMBOL(__start_pci_fixups_enable),
> > +	VMLINUX_SYMBOL(__end_pci_fixups_enable),
> > +	VMLINUX_SYMBOL(__start_pci_fixups_resume),
> > +	VMLINUX_SYMBOL(__end_pci_fixups_resume),
> > +	VMLINUX_SYMBOL(__start_pci_fixups_resume_early),
> > +	VMLINUX_SYMBOL(__end_pci_fixups_resume_early),
> > +	VMLINUX_SYMBOL(__start_pci_fixups_suspend),
> > +	VMLINUX_SYMBOL(__end_pci_fixups_suspend),
> > +	VMLINUX_SYMBOL(__start_builtin_fw),
> > +	VMLINUX_SYMBOL(__end_builtin_fw),
> > +	VMLINUX_SYMBOL(__start_rio_route_ops),
> > +	VMLINUX_SYMBOL(__end_rio_route_ops),
> > +	VMLINUX_SYMBOL(__start___ksymtab),
> > +	VMLINUX_SYMBOL(__stop___ksymtab),
> > +	VMLINUX_SYMBOL(__start___ksymtab_gpl),
> > +	VMLINUX_SYMBOL(__stop___ksymtab_gpl),
> > +	VMLINUX_SYMBOL(__start___ksymtab_unused),
> > +	VMLINUX_SYMBOL(__stop___ksymtab_unused),
> > +	VMLINUX_SYMBOL(__start___ksymtab_unused_gpl),
> > +	VMLINUX_SYMBOL(__stop___ksymtab_unused_gpl),
> > +	VMLINUX_SYMBOL(__start___ksymtab_gpl_future),
> > +	VMLINUX_SYMBOL(__stop___ksymtab_gpl_future),
> > +	VMLINUX_SYMBOL(__start___kcrctab),
> > +	VMLINUX_SYMBOL(__stop___kcrctab),
> > +	VMLINUX_SYMBOL(__start___kcrctab_gpl),
> > +	VMLINUX_SYMBOL(__stop___kcrctab_gpl),
> > +	VMLINUX_SYMBOL(__start___kcrctab_unused),
> > +	VMLINUX_SYMBOL(__stop___kcrctab_unused),
> > +	VMLINUX_SYMBOL(__start___kcrctab_unused_gpl),
> > +	VMLINUX_SYMBOL(__stop___kcrctab_unused_gpl),
> > +	VMLINUX_SYMBOL(__start___kcrctab_gpl_future),
> > +	VMLINUX_SYMBOL(__stop___kcrctab_gpl_future),
> > +	VMLINUX_SYMBOL(__start___param),
> > +	VMLINUX_SYMBOL(__stop___param),
> > +	VMLINUX_SYMBOL(__end_rodata),
> > +	VMLINUX_SYMBOL(__security_initcall_start),
> > +	VMLINUX_SYMBOL(__security_initcall_end),
> > +	VMLINUX_SYMBOL(__sched_text_start),
> > +	VMLINUX_SYMBOL(__sched_text_end),
> > +	VMLINUX_SYMBOL(__lock_text_start),
> > +	VMLINUX_SYMBOL(__lock_text_end),
> > +	VMLINUX_SYMBOL(__kprobes_text_start),
> > +	VMLINUX_SYMBOL(__kprobes_text_end),
> > +	VMLINUX_SYMBOL(__irqentry_text_start),
> > +	VMLINUX_SYMBOL(__irqentry_text_end),
> > +	VMLINUX_SYMBOL(__start___verbose_strings),
> > +	VMLINUX_SYMBOL(__stop___verbose_strings),
> > +	VMLINUX_SYMBOL(__start___verbose),
> > +	VMLINUX_SYMBOL(__stop___verbose),
> > +	VMLINUX_SYMBOL(__start___bug_table),
> > +	VMLINUX_SYMBOL(__stop___bug_table),
> > +	VMLINUX_SYMBOL(__tracedata_start),
> > +	VMLINUX_SYMBOL(__tracedata_end),
> > +	VMLINUX_SYMBOL(__start_notes),
> > +	VMLINUX_SYMBOL(__stop_notes),
> > +	VMLINUX_SYMBOL(__early_initcall_end),
> > +	VMLINUX_SYMBOL(__per_cpu_start),
> > +	VMLINUX_SYMBOL(__per_cpu_end)
> > +};
> 
> I like how you try to solve this at symbol table generation 
> time.
> 
> Instead of this hardcoded table, couldnt we use some more 
> flexible and more future-proof method? Such as ordering 
> same-address symbols by underscores:
> 
>   [same address]
> 
>   non-underscore symbols first        XYZ
>   single-undescroe symbols second     _XYZ
>   double-underscore symbols third     __XYZ
> 
> that way the scheme would be more or less self-maintaining as an 
> underscore already carries a "this is a special, internal 
> symbol" notion.
> 
> 	Ingo

Thanks Lai!
I didn't know how to solve properly this problem.
But indeed, the underscore priority sorted symbols seems to me a good
and scalable idea.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/