linux-trace-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Yordan Karadzhov <y.karadz@gmail.com>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: linux-trace-devel@vger.kernel.org
Subject: Re: [PATCH v2 3/9] kernel-shark: Fix KS_DEFINE_PLUGIN_CONTEXT macro
Date: Mon, 10 May 2021 23:18:38 +0300	[thread overview]
Message-ID: <0dc10813-0fe9-2506-4c32-028892e17082@gmail.com> (raw)
In-Reply-To: <20210510152349.3de97b82@gandalf.local.home>



On 10.05.21 г. 22:23, Steven Rostedt wrote:
> On Mon, 10 May 2021 21:50:57 +0300
> Yordan Karadzhov <y.karadz@gmail.com> wrote:
> 
>> On 10.05.21 г. 21:25, Steven Rostedt wrote:
>>> On Mon, 10 May 2021 14:53:08 +0300
>>> Yordan Karadzhov <y.karadz@gmail.com> wrote:
>>>    
>>>>> Can you show me how you see this error, because this solution does not make
>>>>> any sense.
>>>>
>>>> The problem is that some plugins can build from multiple source files.
>>>> For example in the case when part of the plugin is written in C and
>>>> another part in C++. In those cases we cannot have the functions being
>>>> static.
>>>
>>> So it's because its a mixture of C and C++ code? And you can't make them
>>> static?
>>
>> No, this has no direct connection with C++. The static functions are the
>> same in C++.
>> It became an issue if you have multiple source files. If this is the
>> case you have to put the macro in a header file, so that you can use the
>> functions defined in the macro in all your source files. But this does
>> not work, because the macro defines some global variables as well. To
>> solve this I defined second macro to be used only in the header, but
>> then the functions can't be static.
>>
> 
> Still does not make sense. Obviously, I'm missing something to connect the
> dots.
> 
> Can you show me the error message that you are fixing.

It is not an error message. It is a malfunctioning. Say you have a macro 
that looks like this:

#define KS_DEFINE_PLUGIN_CONTEXT(type)		\
static type **__context_handler;		\
static ssize_t __n_streams = -1;		\
static inline type *__get_context(int sd)	\
{						\
	if (sd < 0 || sd >= __n_streams)	\
		return NULL;			\
	return __context_handler[sd];		\
}
						\
and you want to use __get_context(int sd) in different source files.
If you just place the macro in a header, each source file will have its 
own "static __context_handler" variable which is not what I would like.
So this is a malfunctioning.

I can solve this problem by having a second macro that only declares 
"type *__get_context(int sd)" and place this macro in the header, while 
keeping the original macro in one of the source files, but then 
__get_context() cannot be static. If I go this way and I have multiple 
plugins using the same macro I will have multiple versions of the same 
function and this time the dynamic linking will start to misbehave.

Does that make sense?

Thanks!
Yordan

> 
> Thanks,
> 
> -- Steve
> 

  reply	other threads:[~2021-05-10 20:18 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-04-28 13:47 [PATCH v2 0/9] (Not so) Minor fixes toward KS 2.0 Yordan Karadzhov (VMware)
2021-04-28 13:47 ` [PATCH v2 1/9] kernel-shark: Fix the build for 32b systems Yordan Karadzhov (VMware)
2021-04-28 13:47 ` [PATCH v2 2/9] kernel-shark: Add "cron" job to workflows Yordan Karadzhov (VMware)
2021-04-28 13:47 ` [PATCH v2 3/9] kernel-shark: Fix KS_DEFINE_PLUGIN_CONTEXT macro Yordan Karadzhov (VMware)
2021-05-06 18:11   ` Steven Rostedt
2021-05-10 11:53     ` Yordan Karadzhov
2021-05-10 18:25       ` Steven Rostedt
2021-05-10 18:50         ` Yordan Karadzhov
2021-05-10 19:23           ` Steven Rostedt
2021-05-10 20:18             ` Yordan Karadzhov [this message]
2021-05-10 20:34             ` Yordan Karadzhov
2021-05-10 21:02               ` Steven Rostedt
2021-05-11 13:30                 ` Yordan Karadzhov
2021-04-28 13:47 ` [PATCH v2 4/9] kernel-shark: Add cleanup of all plugin contexts Yordan Karadzhov (VMware)
2021-04-28 13:47 ` [PATCH v2 5/9] kernel-shark: Fix memory leak in "sched events" plugin Yordan Karadzhov (VMware)
2021-04-28 13:47 ` [PATCH v2 6/9] kernel-shark: Disable the pop-up offset dialog Yordan Karadzhov (VMware)
2021-04-28 13:47 ` [PATCH v2 7/9] kernel-shark: Remove kvm_combo from the list of default plugins Yordan Karadzhov (VMware)
2021-05-06 18:25   ` Steven Rostedt
2021-05-10 12:15     ` Yordan Karadzhov
2021-04-28 13:47 ` [PATCH v2 8/9] kernel-shark: Remove debugging print out from plugins Yordan Karadzhov (VMware)
2021-04-28 13:47 ` [PATCH v2 9/9] kernel-shark: Hide all plugin internals Yordan Karadzhov (VMware)
2021-05-06 18:26 ` [PATCH v2 0/9] (Not so) Minor fixes toward KS 2.0 Steven Rostedt

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=0dc10813-0fe9-2506-4c32-028892e17082@gmail.com \
    --to=y.karadz@gmail.com \
    --cc=linux-trace-devel@vger.kernel.org \
    --cc=rostedt@goodmis.org \
    /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).