linux-kernel - Re: [PATCH 4/5] fbcon: Avoid hard-coding built-in font charcount

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20201028053034.GA1205431@PWN>
Date:   Wed, 28 Oct 2020 01:30:34 -0400
From:   Peilin Ye <yepeilin.cs@...il.com>
To:     Daniel Vetter <daniel@...ll.ch>
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>,
        Jiri Slaby <jirislaby@...nel.org>,
        dri-devel@...ts.freedesktop.org, linux-fbdev@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 4/5] fbcon: Avoid hard-coding built-in font charcount

On Tue, Oct 27, 2020 at 08:13:53PM +0100, Daniel Vetter wrote:
> On Tue, Oct 27, 2020 at 12:37:29PM -0400, Peilin Ye wrote:
> > fbcon_startup() and fbcon_init() are hard-coding the number of characters
> > of our built-in fonts as 256. Recently, we included that information in
> > our kernel font descriptor `struct font_desc`, so use `font->charcount`
> > instead of a hard-coded value.
> > 
> > This patch depends on patch "Fonts: Add charcount field to font_desc".
> > 
> > Signed-off-by: Peilin Ye <yepeilin.cs@...il.com>
> 
> So I think this is correct, but it also doesn't do a hole lot yet. fbcon.c
> still has tons of hard-coded 256 all over, and if (p->userfont).
> 
> I think if we instead set vc->vc_font.charcount both in fbcon_init and in
> fbcon_do_set_font (probably just replace the userfont parameter with
> font_charcount for now), then we could replace these all with
> vc->vc_font.charcount. And the code would already improve quite a bit I
> think.
> 
> With just this change here I think we have even more inconsistency, since
> for built-in fonts vc->vc_font.charcount is now set correctly, but for
> userfonts we need to instead look at FNTCHARCNT(vc->vc_font.data).

You are right, let's remove FNTCHARCNT() altogether. fbcon_do_set_font()
still needs a userfont parameter for refcount handling, I'll just add a
charcount parameter to it.

Peilin