linux-kernel - Re: [PATCH v2 next] ACPI: APEI: EINJ: prevent memory corruption in error_type

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <757b5117-a865-4d31-b566-248048f87b29@suswa.mountain>
Date: Wed, 25 Jun 2025 21:19:24 +0300
From: Dan Carpenter <dan.carpenter@...aro.org>
To: Uwe Kleine-König <u.kleine-koenig@...libre.com>
Cc: Zaid Alali <zaidal@...amperecomputing.com>,
	"Rafael J. Wysocki" <rafael@...nel.org>,
	Len Brown <lenb@...nel.org>, James Morse <james.morse@....com>,
	Tony Luck <tony.luck@...el.com>, Borislav Petkov <bp@...en8.de>,
	Ira Weiny <ira.weiny@...el.com>,
	Jonathan Cameron <Jonathan.Cameron@...wei.com>,
	Dan Williams <dan.j.williams@...el.com>,
	Al Viro <viro@...iv.linux.org.uk>,
	Sudeep Holla <sudeep.holla@....com>, linux-acpi@...r.kernel.org,
	linux-kernel@...r.kernel.org, kernel-janitors@...r.kernel.org
Subject: Re: [PATCH v2 next] ACPI: APEI: EINJ: prevent memory corruption in
 error_type_set()

On Wed, Jun 25, 2025 at 08:04:14PM +0200, Uwe Kleine-König wrote:
> Hello Dan,
> 
> On Wed, Jun 25, 2025 at 11:34:38AM -0500, Dan Carpenter wrote:
> > The "einj_buf" buffer is 32 chars.  Verify that "count" is not too large
> > for that.  Also leave the last character as a NUL terminator to ensure
> > the string is properly terminated.
> > 
> > Fixes: 0c6176e1e186 ("ACPI: APEI: EINJ: Enable the discovery of EINJv2 capabilities")
> > Signed-off-by: Dan Carpenter <dan.carpenter@...aro.org>
> > ---
> > v2: I messed up the sizeof() calculation in the copy_from_user() and I put
> >     the parentheses in the wrong place in v1.
> > 
> >  drivers/acpi/apei/einj-core.c | 5 ++++-
> >  1 file changed, 4 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/acpi/apei/einj-core.c b/drivers/acpi/apei/einj-core.c
> > index d6d7e36e3647..2206cbbdccfa 100644
> > --- a/drivers/acpi/apei/einj-core.c
> > +++ b/drivers/acpi/apei/einj-core.c
> > @@ -826,8 +826,11 @@ static ssize_t error_type_set(struct file *file, const char __user *buf,
> >  	int rc;
> >  	u64 val;
> >  
> > +	if (count > sizeof(einj_buf))
> > +		return -EINVAL;
> > +
> >  	memset(einj_buf, 0, sizeof(einj_buf));
> > -	if (copy_from_user(einj_buf, buf, count))
> 
> I would add a comment here to explain the -1. It's in the commit log,
> but that doesn't help a future reader.
> 

Sure.  I can add a comment.

	/* Leave the last character for the NUL terminator */

> > +	if (copy_from_user(einj_buf, buf, min(count, sizeof(einj_buf) - 1)))
> >  		return -EFAULT;
> >  
> >  	if (strncmp(einj_buf, "V2_", 3) == 0) {
> 
> Further down there is:
> 
> 	return count;
> 
> Is that still correct given that you read less now?

Yep.  If we don't return count userspace will try again, right?

regards,
dan carpenter