Re: [PATCH v2 1/2] jump_label: Explicitly disable jump labels in __init code

From: Ingo Molnar
Date: Sun Feb 18 2018 - 08:05:35 EST



* Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:

> On Sat, 17 Feb 2018, Josh Poimboeuf wrote:
> > On Sat, Feb 17, 2018 at 11:38:48AM +0100, Ingo Molnar wrote:
> > >
> > > * Josh Poimboeuf <jpoimboe@xxxxxxxxxx> wrote:
> > >
> > > > +/* Disable any jump label entries in __init code */
> > > > +void __init jump_label_invalidate_init(void)
> > > > +{
> > > > + struct jump_entry *iter_start = __start___jump_table;
> > > > + struct jump_entry *iter_stop = __stop___jump_table;
> > > > + struct jump_entry *iter;
> > > > +
> > > > + for (iter = iter_start; iter < iter_stop; iter++)
> > > > + if (iter->code >= (unsigned long)_sinittext &&
> > > > + iter->code < (unsigned long)_einittext)
> > > > + iter->code = 0;
> > > > +}
> > > > +
> > > > +/* Disable any jump label entries in module init code */
> > > > static void jump_label_invalidate_module_init(struct module *mod)
> > > > {
> > > > struct jump_entry *iter_start = mod->jump_entries;
> > > > struct jump_entry *iter_stop = iter_start + mod->num_jump_entries;
> > > > struct jump_entry *iter;
> > > >
> > > > - for (iter = iter_start; iter < iter_stop; iter++) {
> > > > + for (iter = iter_start; iter < iter_stop; iter++)
> > > > if (within_module_init(iter->code, mod))
> > > > iter->code = 0;
> > > > - }
> > >
> > > Why did you remove the curly braces? They are canonical kernel style for
> > > multi-line statements.
> >
> > Personally I prefer the more compact version, but I have no problem
> > changing it.
>
> Yes, it's certainly a matter of taste. Here is the reason why myself and
> others prefer the version with braces:
>
> https://marc.info/?l=linux-kernel&m=148467980905537&w=2

Easier visual parsing is indeed one of the primary reasons, but there's
two other reasons as well:

2) code robustness

For example:

for (i = 0; i < 10; i++)
if (foo)
bar(i);
baz(i);

Is probably buggy code, although technically it's valid syntax and will compile
just fine.

If all multi-line statements have curly braces then this type of bug cannot occur:

for (i = 0; i < 10; i++) {
if (foo)
bar(i);
baz(i);
}

3) style consistency

Nothing is worse than randomly inconsistent coding style, and in arch/x86/ and
core kernel code using curly braces is certainly the dominant style:

# multi-line C statements without braces:
$ find arch/x86/ kernel mm -name "*.[ch]" | xargs awk '/for \(.*[^{]$/ { line1=$0;
f=1; next } f == 1 && /if \(.*[^{]$/ { f=0; line2=$0; i=1; next } i == 1 { i=0;
line3=$0; j=1; next } j == 1 && /^$/{j=0; print line1; print line2; print line3;
print; next} { f=0; i=0; j=0; }' |grep 'for (' |wc -l

55

# multi-line C statements with braces:
$ find arch/x86 kernel mm -name "*.[ch]" | xargs awk '/for \(.*{$/ { line1=$0; f=1;
next } f == 1 && /if \(.*[^{]$/ { f=0; line2=$0; i=1; next } i == 1 { i=0;
line3=$0; j=1; next } j == 1 && /}/{j=0; print line1; print line2; print line3;
print } { f=0; i=0; j=0; }' |grep 'for (' |wc -l

116

Thanks,

Ingo