mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Chris Gellermann <christian.gellermann@codasip.com>
To: bongsu.jeon@samsung.com, david@ixit.cz, netdev@vger.kernel.org
Cc: oe-linux-nfc@lists.linux.dev, linux-kernel@vger.kernel.org,
	Chris Gellermann <christian.gellermann@codasip.com>
Subject: [PATCH] nfc: virtual_ncidev: Add missing ioctl compat handler
Date: Fri,  4 Sep 2026 18:42:52 +0200	[thread overview]
Message-ID: <20260904164252.18351-1-christian.gellermann@codasip.com> (raw)

The compat handler for ioctls to the virtual nci device is missing. So,
nci-specific ioctls of a compat task return with -1 and errno set to
ENOTTY. Add a handler.

The handling of an ioctl() call of a compat task to get the index of
virtual nci device (IOCTL_GET_NCIDEV_IDX) lands in the default case of
the ioctl compat handler (see fs/ioctl.c):

COMPAT_SYSCALL_DEFINE3(ioctl, ...)
{
	...
	default:
       		error = do_vfs_ioctl(fd_file(f), fd, cmd, ...);
		if (error != -ENOIOCTLCMD)
			break;

		if (fd_file(f)->f_op->compat_ioctl)
			error = fd_file(f)->f_op->compat_ioctl(fd_file(f), cmd, arg);
		if (error == -ENOIOCTLCMD)
			error = -ENOTTY;
	...
}

There, do_vfs_ioctl() returns -ENOIOCTLCMD and compat_ioctl is not
set for virtual_ncidev_fops, i.e. f_op->compat_ioctl == NULL. So, the
ioctl() syscall returns with -1 and errno set to ENOTTY to the compat
task.

To fix this, use the compat_ptr_ioctl helper for compat handling here.
It shall be used for ioctls that "either ignore the argument or pass a
pointer to a compatible data type". The driver's sole ioctl takes a user
void pointer and copies nfc_dev->idx to it, a 4-byte integer across all
ABIs.

This issue has been found by running the nci_dev kernel selftest as
rv64 binary on top of a CHERI kernel, where the ioctl() ends up in
the ioctl compat handler, similar to a 32-bit application on top of a
64-bit kernel.

Fixes: e624e6c3e777 ("nfc: Add a virtual nci device driver")
Signed-off-by: Chris Gellermann <christian.gellermann@codasip.com>
---
 drivers/nfc/virtual_ncidev.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/nfc/virtual_ncidev.c b/drivers/nfc/virtual_ncidev.c
index 8eeb447ac96e..e51c647b27eb 100644
--- a/drivers/nfc/virtual_ncidev.c
+++ b/drivers/nfc/virtual_ncidev.c
@@ -195,7 +195,8 @@ static const struct file_operations virtual_ncidev_fops = {
 	.write = virtual_ncidev_write,
 	.open = virtual_ncidev_open,
 	.release = virtual_ncidev_close,
-	.unlocked_ioctl = virtual_ncidev_ioctl
+	.unlocked_ioctl = virtual_ncidev_ioctl,
+	.compat_ioctl = compat_ptr_ioctl,
 };
 
 static struct miscdevice miscdev = {
-- 
2.47.3


             reply	other threads:[~2026-09-04 16:43 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-04 16:42 Chris Gellermann [this message]
2026-09-08  9:34 ` Simon Horman

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=20260904164252.18351-1-christian.gellermann@codasip.com \
    --to=christian.gellermann@codasip.com \
    --cc=bongsu.jeon@samsung.com \
    --cc=david@ixit.cz \
    --cc=linux-kernel@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=oe-linux-nfc@lists.linux.dev \
    /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®