From: Kees Cook <keescook@chromium.org>
To: Romain Perier <romain.perier@gmail.com>
Cc: kernel-hardening@lists.openwall.com
Subject: Re: [PRE-REVIEW PATCH 11/16] treewide: Globally replace tasklet_init() by tasklet_setup()
Date: Thu, 10 Oct 2019 15:30:52 -0700 [thread overview]
Message-ID: <201910101529.D6550C790D@keescook> (raw)
In-Reply-To: <20191001171828.GB2748@debby.home>
On Tue, Oct 01, 2019 at 07:18:28PM +0200, Romain Perier wrote:
> On Mon, Sep 30, 2019 at 03:46:29PM -0700, Kees Cook wrote:
> > On Sun, Sep 29, 2019 at 06:30:23PM +0200, Romain Perier wrote:
> > > This converts all remaining cases of the old tasklet_init() API into
> > > tasklet_setup(), where the callback argument is the structure already
> > > holding the struct tasklet_struct. These should have no behavioral changes,
> > > since they just change which pointer is passed into the callback with
> > > the same available pointers after conversion. Moreover, all callbacks
> > > that were not passing a pointer of structure holding the struct
> > > tasklet_struct has already been converted.
> >
> > Was this done mechanically with Coccinelle or manually? (If done with
> > Coccinelle, please include the script in the commit log.) To land a
> > treewide change like this usually you'll need to separate the mechanical
> > from the manual as Linus likes to run those changes himself sometimes.
>
> Hi,
>
> This was done with both technics mechanically with a "buggy" Coccinelle
> script, after what I have fixed building errors and mismatches (even if it's
> clearly super powerful, it was my first complex cocci script). 80% of trivial
> replacements were done with a Cocci script, the rest was done manually.
> That's complicated to remember which one was mechanically or manually to
> be honnest :=D
>
> What I can propose is the following:
>
> - A commit for trivial tasklet_init() -> tasklet_setup() replacements:
> it would contain basic replacements of the calls "tasklet_init() ->
> tasklet_setup()" and addition of "from_tasklet()" without any other
> changes.
Right -- the manual ones might need to be split up by subsystem or
driver.
> - A second commit for more complicated replacements:
> It would contain replacements of functions that are in different
> modules, or modules that use function pointer for tasklet handlers
> etc... Basically everything that is not covered by the first commit
Same for this if it can't be automated.
> What do you think ?
> Moreover, the cocci script I have used is... ugly... so I don't want to
> see Linus's eyes bleed :=D
Heh. Well, the timer_struct Cocci script was ugly too. The idea is that
maintainers will likely want per-driver patches, so the more you can
automate with a script to put in a single commit for Linus would be
better for your own sanity. :)
> PS: I can try to recover the cocci script in my git repo by using "git
> reflog". And put the cocci script in the first commit (for trivial
> replacements), in the worst case...
Probably it is only needed in the commit log.
--
Kees Cook
next prev parent reply other threads:[~2019-10-10 22:31 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-09-29 16:30 [PRE-REVIEW PATCH 00/16] Modernize the tasklet API Romain Perier
2019-09-29 16:30 ` [PRE-REVIEW PATCH 01/16] tasklet: Prepare to change tasklet callback argument type Romain Perier
2019-09-29 16:30 ` [PRE-REVIEW PATCH 02/16] crypto: ccp - Prepare to use the new tasklet API Romain Perier
2019-09-30 22:35 ` Kees Cook
2019-09-29 16:30 ` [PRE-REVIEW PATCH 03/16] mmc: renesas_sdhi: " Romain Perier
2019-09-29 16:30 ` [PRE-REVIEW PATCH 04/16] net: liquidio: " Romain Perier
2019-09-29 16:30 ` [PRE-REVIEW PATCH 05/16] chelsio: " Romain Perier
2019-09-29 16:30 ` [PRE-REVIEW PATCH 06/16] net: mvpp2: " Romain Perier
2019-09-29 16:30 ` [PRE-REVIEW PATCH 07/16] qed: " Romain Perier
2019-09-29 16:30 ` [PRE-REVIEW PATCH 08/16] isdn: " Romain Perier
2019-09-29 16:30 ` [PRE-REVIEW PATCH 09/16] scsi: pm8001: " Romain Perier
2019-09-29 16:30 ` [PRE-REVIEW PATCH 10/16] scsi: pmcraid: " Romain Perier
2019-09-29 16:30 ` [PRE-REVIEW PATCH 11/16] treewide: Globally replace tasklet_init() by tasklet_setup() Romain Perier
2019-09-30 22:46 ` Kees Cook
2019-10-01 17:18 ` Romain Perier
2019-10-10 22:30 ` Kees Cook [this message]
2019-09-29 16:30 ` [PRE-REVIEW PATCH 12/16] tasklet: Pass tasklet_struct pointer as .data in DECLARE_TASKLET Romain Perier
2019-09-30 22:44 ` Kees Cook
2019-09-29 16:30 ` [PRE-REVIEW PATCH 13/16] tasklet: Pass tasklet_struct pointer to callbacks unconditionally Romain Perier
2019-09-30 22:49 ` Kees Cook
2019-09-29 16:30 ` [PRE-REVIEW PATCH 14/16] tasklet: Remove the data argument from DECLARE_TASKLET() macros Romain Perier
2019-09-30 22:50 ` Kees Cook
2019-09-29 16:30 ` [PRE-REVIEW PATCH 15/16] tasklet: convert callbacks prototype for using struct tasklet_struct * arguments Romain Perier
2019-09-29 16:30 ` [PRE-REVIEW PATCH 16/16] tasklet: Add the new initialization function permanently Romain Perier
2019-09-30 22:52 ` Kees Cook
2019-10-01 17:34 ` Romain Perier
2019-09-30 23:06 ` [PRE-REVIEW PATCH 00/16] Modernize the tasklet API Kees Cook
2019-10-01 17:47 ` Romain Perier
2019-10-10 22:34 ` Kees Cook
2019-10-30 8:20 ` Allen
2019-11-07 7:29 ` Romain Perier
2019-11-07 21:22 ` Kees Cook
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=201910101529.D6550C790D@keescook \
--to=keescook@chromium.org \
--cc=kernel-hardening@lists.openwall.com \
--cc=romain.perier@gmail.com \
/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).