Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Tue, 21 May 2019 22:43:11 -0400 (EDT)
From:   Nicolas Pitre <nico@fluxnic.net>
To:     Gen Zhang <blackgod016574@gmail.com>
cc:     linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2] vt: Fix a missing-check bug in drivers/tty/vt/vt.c
In-Reply-To: <20190521073901.GF5263@zhanggen-UX430UQ>
Message-ID: <nycvar.YSQ.7.76.1905212218090.1558@knanqh.ubzr>
References: <20190521022940.GA4858@zhanggen-UX430UQ> <nycvar.YSQ.7.76.1905202242410.1558@knanqh.ubzr> <20190521030905.GB5263@zhanggen-UX430UQ> <nycvar.YSQ.7.76.1905202323290.1558@knanqh.ubzr> <20190521040019.GD5263@zhanggen-UX430UQ>
 <nycvar.YSQ.7.76.1905210022050.1558@knanqh.ubzr> <20190521073901.GF5263@zhanggen-UX430UQ>
User-Agent: Alpine 2.21 (LFD 202 2017-01-01)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, 21 May 2019, Gen Zhang wrote:

> On Tue, May 21, 2019 at 12:30:38AM -0400, Nicolas Pitre wrote:
> > Now imagine that MIN_NR_CONSOLES is defined to 10 instead of 1.
> > 
> > What happens with allocated memory if the err_vc condition is met on the 
> > 5th loop?
> Yes, vc->vc_screenbuf from the last loop is still not freed, right? I
> don't have idea to solve this one. Could please give some advice? Since
> we have to consider the err_vc condition.
> 
> > If err_vc_screenbuf condition is encountered on the 5th loop (curcons = 
> > 4), what is the value of vc_cons[4].d? Isn't it the same as vc that you 
> > just freed?
> > 
> > 
> > Nicolas
> Thanks for your explaination! You mean a double free situation may 
> happen, right? But in vc_allocate() there is also such a kfree(vc) and 
> vc_cons[currcons].d = NULL operation. This situation is really confusing
> me.

What you could do is something that looks like:

	for (currcons = 0; currcons < MIN_NR_CONSOLES; currcons++) {
		vc_cons[currcons].d = vc = kzalloc(...);
		if (!vc)
			goto fail1;
		...
		vc->vc_screenbuf = kzalloc(...);
		if (!vc->vc_screenbuf)
			goto fail2;
		...

	return 0;

	/* free already allocated memory on error */
fail1:
	while (curcons > 0) {
		curcons--;
		kfree(vc_cons[currcons].d->vc_screenbuf);
fail2:
		kfree(vc_cons[currcons].d);
		vc_cons[currcons].d = NULL;
	}
	console_unlock();
	return -ENOMEM;


Nicolas