[PATCH] mtd: nand: don't leak buffers when ->scan_bbt() fails

[Brian Norris]

On Mon, May 01, 2017 at 09:22:03PM -0300, Ezequiel Garcia wrote:
> On 1 May 2017 at 21:04, Brian Norris <computersforpeace at gmail.com> wrote:
> > This bug seems to have been here forever, although we came close to
> > fixing all of them in [1]!
> >
> > [1] 11eaf6df1cce ("mtd: nand: Remove BUG() abuse in nand_scan_tail")
> >
> 
> Well, we came even closer. See [1] for a patch that fixes this by cleaning
> BBT init and release.

That's a different bug(fix), no?

> Back then Boris suggested on IRC to wait for Peter Pan's BBT, which would
> address this too.

Haha, well bad strategy I guess :)

> It's interesting to note that: (1) the patch never got any formal feedback,
> despite it was a fix; and (2) Peter Pan's work is still under development.

Yes, well you kinda nacked yourself, so (1) isn't that surprising. And
(2) is well...no comment.

> Guess I should have pressed the upstream buttons harder :-)

I suppose.

I'd be happy to see you resubmit, since I definitely would prioritize
bugfixes over years-long refactoring. But I guess I'd defer to Boris, et
al, who are paying much more attention to NAND stuff these days than I
am. I just noticed this when bugfixing some things I noticed in Boris's
pull request.

Brian

> [1] http://lists.infradead.org/pipermail/linux-mtd/2016-April/066834.html,
> [2] http://lists.infradead.org/pipermail/linux-mtd/2016-April/066835.html
> -- 
> Ezequiel García, VanguardiaSur
> www.vanguardiasur.com.ar