mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Syed Tayyab Farooq <syedtayyabfarooq08@gmail.com>
To: Helge Deller <deller@gmx.de>,
	Thomas Zimmermann <tzimmermann@suse.de>,
	Simona Vetter <simona@ffwll.ch>, Andrew Morton <akpm@osdl.org>,
	Antonino Daplas <adaplas@pol.net>,
	dri-devel@lists.freedesktop.org (open list:FRAMEBUFFER CONSOLE),
	linux-fbdev@vger.kernel.org (open list:FRAMEBUFFER CONSOLE),
	linux-kernel@vger.kernel.org (open list)
Cc: Syed Tayyab Farooq <syedtayyabfarooq08@gmail.com>,
	syzbot+bf4020519fbd94b2355d@syzkaller.appspotmail.com
Subject: [PATCH] fbdev: fbcon: Fix KASAN slab-out-of-bounds read in rotated font buffer
Date: Wed, 23 Sep 2026 11:08:07 +0500	[thread overview]
Message-ID: <20260923060924.45742-1-syedtayyabfarooq08@gmail.com> (raw)

The root cause of this KASAN slab-out-of-bounds read is that fbcon fails to
reallocate the rotated font buffer (par->rotated.buf) when the font
character count changes but the font dimensions (width and height) remain
the same.

When a user changes the font via ioctl(KDFONTOP), fbcon_do_set_font() is
called. If the user first sets a 256-character font, enables rotation, and
then sets a 512-character font with the exact same width and height, resize
evaluates to false. The function then directly calls update_screen(vc).
Notice that fbcon_do_set_font() never calls par->bitops->rotate_font() to
update the rotated font buffer for the new font. As a result,
par->rotated.buf remains sized for the old 256-character font.

When update_screen(vc) is called, it eventually reaches cw_putcs_aligned(),
which attempts to read the glyphs from the rotated buffer. Since
vc->vc_font.charcount is now 512, vc->vc_hi_font_mask is set, making
charmask equal to 0x1ff (511). If the text buffer contains a character with
an index >= 256, src will point out of bounds of the old par->rotated.buf
(which only holds 256 glyphs). The subsequent read in
__fb_pad_aligned_buffer() triggers the KASAN slab-out-of-bounds crash.

To fix this, fbcon_do_set_font() must call par->bitops->rotate_font() to
update the rotated font buffer before calling vc_resize() or
update_screen(). If vc_resize() fails and the old font is restored in the
error path, rotate_font() must be called again to restore the rotated
buffer.

This pattern of updating the rotated font buffer was previously established
in fbcon_modechanged() by commit ef7656e85f1a ("fbdev: Fix out-of-bounds
access when rotating console after font resize"). The fix applies this
exact idiom to fbcon_do_set_font(). By placing it before vc_resize() and
update_screen(), we ensure the buffer is correctly sized before any
rendering occurs. Checking con_is_visible(vc) is also correct here, as
hidden consoles will have their rotation state updated later via
fbcon_switch() when they become visible.

Furthermore, by calling rotate_font() again in the err_out path, we ensure
that if vc_resize() fails, the rotated buffer is restored to match the old
vc->vc_font. Without this, restoring a 512-character font to a
256-character font on failure would leave the rotated buffer sized for 512
characters, which wastes memory; or worse, restoring a 256-character font
to a 512-character font would leave the buffer too small, causing an
out-of-bounds read later.

Fixes: e4fc27618b75 ("[PATCH] fbcon: Console Rotation - Prepare fbcon for console rotation")
Assisted-by: Gemini:gemini-3.6-flash Gemini:gemini-3.1-pro-preview syzbot
Reported-by: syzbot+bf4020519fbd94b2355d@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=bf4020519fbd94b2355d
Link: https://syzkaller.appspot.com/ai_job?id=59f0d7d9-607c-4497-b2e2-0b16693f9502
Signed-off-by: Syed Tayyab Farooq <syedtayyabfarooq08@gmail.com>
---
 drivers/video/fbdev/core/fbcon.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/drivers/video/fbdev/core/fbcon.c b/drivers/video/fbdev/core/fbcon.c
index 01715873ea49..158b2f91d68e 100644
--- a/drivers/video/fbdev/core/fbcon.c
+++ b/drivers/video/fbdev/core/fbcon.c
@@ -2432,6 +2432,13 @@ static int fbcon_do_set_font(struct vc_data *vc, int w, int h, int charcount,
 	else if (!vc->vc_hi_font_mask && charcount == 512)
 		set_vc_hi_font(vc, true);
 
+	if (con_is_visible(vc)) {
+		if (par->bitops->rotate_font && par->bitops->rotate_font(info, vc)) {
+			par->rotate = FB_ROTATE_UR;
+			set_blitting_type(vc, info);
+		}
+	}
+
 	if (resize) {
 		int cols, rows;
 
@@ -2466,6 +2473,13 @@ static int fbcon_do_set_font(struct vc_data *vc, int w, int h, int charcount,
 	else if (!old_hi_font_mask && vc->vc_hi_font_mask)
 		set_vc_hi_font(vc, false);
 
+	if (con_is_visible(vc)) {
+		if (par->bitops->rotate_font && par->bitops->rotate_font(info, vc)) {
+			par->rotate = FB_ROTATE_UR;
+			set_blitting_type(vc, info);
+		}
+	}
+
 	font_data_put(data);
 
 	return ret;
-- 
2.43.0


             reply	other threads:[~2026-09-23  6:10 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-23  6:08 Syed Tayyab Farooq [this message]
2026-11-11 11:30 ` Syed Tayyab Farooq
2026-11-18 21:33 ` Helge Deller

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=20260923060924.45742-1-syedtayyabfarooq08@gmail.com \
    --to=syedtayyabfarooq08@gmail.com \
    --cc=adaplas@pol.net \
    --cc=akpm@osdl.org \
    --cc=deller@gmx.de \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=linux-fbdev@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=simona@ffwll.ch \
    --cc=syzbot+bf4020519fbd94b2355d@syzkaller.appspotmail.com \
    --cc=tzimmermann@suse.de \
    /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

all inboxes | Powered by JetHome®