On Thu 2021-02-18 12:41:39, Chris Down wrote:
> Petr Mladek writes:
> > > - See if it's safe to pass a printk_fmt_sec to seq_file instead of a
> > > module
> >
> > Also it might be needed to store the pointer to struct module.
>
> You mean, have a `struct module` entry for this? I somewhat suspect that
> module.c maintainers are not likely to be happy about injecting non-generic
> code into there if it's possible to be avoided, but maybe I'm
> misunderstanding?
Yes, I suggest to store the pointer into struct module. It includes
many external entries. It is similar to struct task_struct.
I am active also in the kernel livepatch subsystem. We have added
there three values:
#ifdef CONFIG_LIVEPATCH
bool klp; /* Is this a livepatch module? */
bool klp_alive;
/* Elf information */
struct klp_modinfo *klp_info;
#endif
Many other subsystems have their stuff there, for example:
#ifdef CONFIG_TRACING
unsigned int num_trace_bprintk_fmt;
const char **trace_bprintk_fmt_start;
#endif
#ifdef CONFIG_EVENT_TRACING
struct trace_event_call **trace_events;
unsigned int num_trace_events;
struct trace_eval_map **trace_evals;
unsigned int num_trace_evals;
#endif
#ifdef CONFIG_FTRACE_MCOUNT_RECORD
unsigned int num_ftrace_callsites;
unsigned long *ftrace_callsites;
#endif
#ifdef CONFIG_KPROBES
void *kprobes_text_start;
unsigned int kprobes_text_size;
unsigned long *kprobe_blacklist;
unsigned int num_kprobe_blacklist;
#endif
BTW: Jessica originally worked on the kernel livepatching.
She became module loader code maintainer because we needed
even more hacks there and the original maintainer got
busy with other stuff at that time ;-)
I am pretty sure that she would accept it. We need a per-module
value. It is not necessary to maintain separate global list/hash
table just to store these values.
Best Regards,
Petr
