linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Petr Mladek <pmladek@suse.com>
To: jim.cromie@gmail.com
Cc: Jason Baron <jbaron@akamai.com>,
	LKML <linux-kernel@vger.kernel.org>,
	akpm@linuxfoundation.org, Greg KH <gregkh@linuxfoundation.org>,
	Rasmus Villemoes <linux@rasmusvillemoes.dk>
Subject: Re: [PATCH v2 13/24] dyndbg: combine flags & mask into a struct, use that
Date: Tue, 16 Jun 2020 13:34:43 +0200	[thread overview]
Message-ID: <20200616113443.GB2238@alley> (raw)
In-Reply-To: <CAJfuBxyTsrsyMZFhET3yxM1APobY98ykBLuQ2LEhHKsOYtyjEg@mail.gmail.com>

On Mon 2020-06-15 23:47:26, jim.cromie@gmail.com wrote:
> On Mon, Jun 15, 2020 at 9:14 AM Petr Mladek <pmladek@suse.com> wrote:
> >
> > On Sat 2020-06-13 09:57:27, Jim Cromie wrote:
> > > combine flags & mask into a struct, and replace those 2 parameters in
> > > 3 functions: ddebug_change, ddebug_parse_flags, ddebug_read_flags,
> > > altering the derefs in them accordingly.
> > >
> > > This simplifies the 3 function sigs, preparing for more changes.
> > > We dont yet need mask from ddebug_read_flags, but will soon.
> > > ---
> > >  lib/dynamic_debug.c | 46 +++++++++++++++++++++++----------------------
> > >  1 file changed, 24 insertions(+), 22 deletions(-)
> > >
> > > diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c
> > > index 93c627e9c094..8dc073a6e8a4 100644
> > > --- a/lib/dynamic_debug.c
> > > +++ b/lib/dynamic_debug.c
> > > +struct flagsettings {
> > > +     unsigned int flags;
> > > +     unsigned int mask;
> > > +};
> >
> > static int ddebug_change(const struct ddebug_query *query,
> > > -                     unsigned int pflags, unsigned int mask)
> > > +                      struct flagsettings *mods)
> >
> > > -static int ddebug_read_flags(const char *str, unsigned int *flags)
> > > +static int ddebug_read_flags(const char *str, struct flagsettings *f)
> >
> > > -static int ddebug_parse_flags(const char *str, unsigned int *flagsp,
> > > -                            unsigned int *maskp)
> > > +static int ddebug_parse_flags(const char *str, struct flagsettings *mods)
> >
> > What "mods" stands for, please?
> >
> modifying_flags, or modifiers.
> the original flags & mask bundled together
> ie the pfmlt in
>    echo +pfmlt > control

Honestly, storing flags and mask is a hack that makes the code
tricky like hell.

IMHO, it would be much easier to define something like:

struct flags_operation {
	unsinged int flags;
	enum flags_operation_type op;
}

Where the opration would be:

enum flags_operation_type {
	DD_FLAGS_SET,		/* for '=' */
	DD_FLAGS_ADD,		/* for '+' */
	DD_FLAGS_DEL,		/* for '-' */
	DD_FLAGS_FILTER_MATCH,
	DD_FLAGS_FILTER_NON_MATCH,
};

Then you could define

     struct flags_operation fop_change;
     struct flags_operation fop_filter;

Then you could do in ddebug_change():

	if (fop_filter) {
		switch(fop_filter->op) {
			case DD_FLAGS_FILTER_MATCH:
				if ((dp->flags & fop_filter->flags) != fop_filter->flags)
					continue;
				break;
			case: DD_FLAGS_FILTER_NON_MATCH:
				if ((dp->flags & fop_filter->flags)
					continue;
				break;
			default:
				// warn error;
		}
	}

	switch (fop_change->op) {
		case DD_FLAGS_SET:
			dp->flags = fop_change->flags;
			break;
		case DD_FLAGS_ADD:
			dp->flags |= fop_change->flags;
			break;
		case DD_FLAGS_DEL:
			dp->flgas &= ^(fop_change->flgas);
			break;
		default:
			// error;
	}


It might be more lines. But the bit operations will become straightforward.
and the code self-explaining,


> does the above help ?
> I could go with modifying_flags
> keep in mind the name has to suit all + - = operations
> 
> I'll review all the new names. I recall you didnt like filterflags either,
> so I wasnt sufficently clear there either.
> Im mulling a better explanation.


The above would make the code manageable. Another question is the user
interface.

I still wonder if it is worth it.
What is the motivation for this fitlering?
Is it requested by users?
Or is it just a prerequisite for the user-specific filters?

We need to be really careful. User interface is hard to change
or remove later.

Best Regards,
Petr

  reply	other threads:[~2020-06-16 11:34 UTC|newest]

Thread overview: 62+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-13 15:57 [PATCH v2 00/24] dynamic_debug cleanups, query features, WIP print-classes Jim Cromie
2020-06-13 15:57 ` [PATCH v2 01/24] dyndbg-docs: eschew file /full/path query in docs Jim Cromie
2020-06-14  6:04   ` Greg KH
2020-06-14 14:24     ` jim.cromie
2020-06-13 15:57 ` [PATCH v2 02/24] dyndbg-docs: initialization is done early, not arch Jim Cromie
2020-06-13 15:57 ` [PATCH v2 03/24] dyndbg: drop obsolete comment on ddebug_proc_open Jim Cromie
2020-06-13 15:57 ` [PATCH v2 04/24] dyndbg: refine debug verbosity; 1 is basic, 2 more chatty Jim Cromie
2020-06-13 15:57 ` [PATCH v2 05/24] dyndbg: rename __verbose section to __dyndbg Jim Cromie
2020-06-13 15:57 ` [PATCH v2 06/24] dyndbg: fix overcounting of ram used by dyndbg Jim Cromie
2020-06-13 15:57 ` [PATCH v2 07/24] dyndbg: fix a BUG_ON in ddebug_describe_flags Jim Cromie
2020-06-15 13:20   ` Petr Mladek
2020-06-15 21:53     ` jim.cromie
2020-06-13 15:57 ` [PATCH v2 08/24] dyndbg: make ddebug_tables list LIFO for add/remove_module Jim Cromie
2020-06-13 15:57 ` [PATCH v2 09/24] dyndbg: add maybe(str,"") macro to reduce code Jim Cromie
2020-06-13 16:14   ` Joe Perches
2020-06-15 20:50     ` jim.cromie
2020-06-15 13:28   ` Petr Mladek
2020-06-15 20:52     ` jim.cromie
2020-06-13 15:57 ` [PATCH v2 10/24] dyndbg: refactor parse_linerange out of ddebug_parse_query Jim Cromie
2020-06-15 13:37   ` Petr Mladek
2020-06-15 22:37     ` jim.cromie
2020-06-13 15:57 ` [PATCH v2 11/24] dyndbg: accept 'file foo.c:func1' and 'file foo.c:10-100' Jim Cromie
2020-06-15 14:46   ` Petr Mladek
2020-06-13 15:57 ` [PATCH v2 12/24] dyndbg: refactor ddebug_read_flags out of ddebug_parse_flags Jim Cromie
2020-06-13 15:57 ` [PATCH v2 13/24] dyndbg: combine flags & mask into a struct, use that Jim Cromie
2020-06-15 15:14   ` Petr Mladek
2020-06-16  5:47     ` jim.cromie
2020-06-16 11:34       ` Petr Mladek [this message]
2020-06-13 15:57 ` [PATCH v2 14/24] dyndbg: add filter parameter to ddebug_parse_flags Jim Cromie
2020-06-15 15:17   ` Petr Mladek
2020-06-13 15:57 ` [PATCH v2 15/24] dyndbg: extend ddebug_parse_flags to accept optional leading filter-flags Jim Cromie
2020-06-15 15:37   ` Petr Mladek
2020-06-13 15:57 ` [PATCH v2 16/24] dyndbg: prefer declarative init in caller, to memset in callee Jim Cromie
2020-06-13 15:57 ` [PATCH v2 17/24] dyndbg: add user-flag, negating-flags, and filtering on flags Jim Cromie
2020-06-16 11:41   ` Petr Mladek
2020-06-13 15:57 ` [PATCH v2 18/24] dyndbg: allow negating flag-chars in modflags Jim Cromie
2020-06-16 11:47   ` Petr Mladek
2020-06-13 15:57 ` [PATCH v2 19/24] dyndbg: accept query terms like module:foo and file=bar Jim Cromie
2020-06-16 11:57   ` Petr Mladek
2020-06-16 20:08     ` jim.cromie
2020-06-17 12:11       ` Petr Mladek
2020-06-17 13:32         ` jim.cromie
2020-06-17 14:01           ` Greg KH
2020-06-13 15:57 ` [PATCH v2 20/24] dyndbg: WIP towards debug-print-class based callsite controls Jim Cromie
2020-06-13 19:22   ` Joe Perches
2020-06-16 13:45   ` Petr Mladek
2020-06-16 21:05     ` Joe Perches
2020-06-17  9:31       ` Daniel Thompson
2020-06-17  9:52         ` Petr Mladek
2020-06-17 13:23           ` jim.cromie
2020-06-18 12:20             ` Petr Mladek
2020-06-16 21:13     ` jim.cromie
2020-06-16 21:25       ` jim.cromie
2020-06-13 15:57 ` [PATCH v2 21/24] dyndbg: adapt header macros to pass print-class Jim Cromie
2020-06-13 17:11   ` kernel test robot
2020-06-13 20:25   ` kernel test robot
2020-06-13 15:57 ` [PATCH v2 22/24] dyndbg: add print-class as trailing number to control output Jim Cromie
2020-06-13 15:57 ` [PATCH v2 23/24] kset-example: add pr_debug()s for easy visibility of its operation Jim Cromie
2020-06-14  6:05   ` Greg KH
2020-06-15 22:18     ` jim.cromie
2020-06-16  6:37       ` Greg KH
2020-06-13 15:57 ` [PATCH v2 24/24] kset-example: use pr_debug_n to create example print-classes Jim Cromie

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200616113443.GB2238@alley \
    --to=pmladek@suse.com \
    --cc=akpm@linuxfoundation.org \
    --cc=gregkh@linuxfoundation.org \
    --cc=jbaron@akamai.com \
    --cc=jim.cromie@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux@rasmusvillemoes.dk \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).