linux-kernel - Re: log spammed with "loading xx failed with error -2" since commit e40ba6d56b [replace call to fw_read_file

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]
Date:	Sun, 28 Feb 2016 21:57:55 +0100
From:	"Luis R. Rodriguez" <mcgrof@...e.com>
To:	Ming Lei <ming.lei@...onical.com>
Cc:	Kees Cook <keescook@...omium.org>,
	"Luis R. Rodriguez" <mcgrof@...nel.org>,
	Heiner Kallweit <hkallweit1@...il.com>,
	James Morris <jmorris@...ei.org>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	Mimi Zohar <zohar@...ux.vnet.ibm.com>,
	LKML <linux-kernel@...r.kernel.org>
Subject: Re: log spammed with "loading xx failed with error -2" since commit
 e40ba6d56b [replace call to fw_read_file_contents() with kernel version]

On Sun, Feb 28, 2016 at 10:10:57PM +0800, Ming Lei wrote:
> On Sun, Feb 28, 2016 at 10:27 AM, Kees Cook <keescook@...omium.org> wrote:
> > On Sat, Feb 27, 2016 at 3:15 PM, Luis R. Rodriguez <mcgrof@...nel.org> wrote:
> >> On Sat, Feb 27, 2016 at 05:49:38PM +0100, Heiner Kallweit wrote:
> >>> Since this commit my system log is spammed with firmware load errors. One example:
> >>>
> >>> loading /lib/firmware/updates/4.5.0-rc5-next-20160226/iwlwifi-3160-16.ucode failed with error -2
> >>> loading /lib/firmware/updates/iwlwifi-3160-16.ucode failed with error -2
> >>> loading /lib/firmware/4.5.0-rc5-next-20160226/iwlwifi-3160-16.ucode failed with error -2
> >>> -> finally load attempt from /lib/firmware/iwlwifi-3160-16.ucode succeeds
> >>>
> >>> Before this commit, when a load from one path in fw_path failed, silently the next path was tried.
> >>> Only if all attempts failed an error message was printed.
> >>> Now for each single failed attempt an error message is printed what doesn't make sense.
> >>
> >> To be clear, this is an issue on linux-next, due to the latest merge
> >> of Mimi's common kernel file loader pulled recently by James.
> >>
> >> Heiner, thanks for the report, this patch fixes that. I'll be submitting that
> >> now.  James, should this go through your tree? Please note I've been meaning to
> >> add myself to MAINTAINERS for FIRMWARE_CLASS as requested by Greg at kernel
> >> summit as I've been helping with cleanup there but I hadn't done so as I had
> >> some pending patches with a full new functionality added. With Mimi's changes
> >> merged on linux-next though and the common kernel file loader now done I can
> >> follow up with my series of changes after this.
> >>
> >>   Luis
> >>
> >> From 92e2bd76bf1abb3ce381b8d54ba5486a295af1a9 Mon Sep 17 00:00:00 2001
> >> From: "Luis R. Rodriguez" <mcgrof@...nel.org>
> >> Date: Sat, 27 Feb 2016 14:58:08 -0800
> >> Subject: [PATCH] firmware: change kernel read fail to dev_dbg()
> >>
> >> When we now use the new kernel_read_file_from_path() we
> >> are reporting a failure when we iterate over all the paths
> >> possible for firmware. Before using kernel_read_file_from_path()
> >> we only reported a failure once we confirmed a file existed
> >> with filp_open() but failed with fw_read_file_contents().
> >>
> >> With kernel_read_file_from_path() both are done for us and
> >> we obviously are now reporting too much information given that
> >> some optional paths will always fail and clutter the logs.
> >>
> >> fw_get_filesystem_firmware() already has a check for failure
> >> and uses an internal flag, FW_OPT_NO_WARN, to let users
> >> warn or not warn. For instance request_firmware_direct()
> >> does not warn as this can be used for optional firmware
> >> as it has no usermode helper fallback. In the future we
> >> may want to change this, given everyone is disabling the
> >> usermode helper anyway now, but for now keep reporting
> >> only as was designed. request_firmware_direct() will
> >> continue to not report errors as it was designed not to.
> >>
> >> Reported-by: Heiner Kallweit <hkallweit1@...il.com>
> >> Cc: Heiner Kallweit <hkallweit1@...il.com>
> >> Cc: Mimi Zohar <zohar@...ux.vnet.ibm.com>
> >> Cc: Kees Cook <keescook@...omium.org>
> >> Signed-off-by: Luis R. Rodriguez <mcgrof@...nel.org>
> >> ---
> >>  drivers/base/firmware_class.c | 2 +-
> >>  1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c
> >> index 1cff832ab74e..b1cf4d61ffc9 100644
> >> --- a/drivers/base/firmware_class.c
> >> +++ b/drivers/base/firmware_class.c
> >> @@ -328,7 +328,7 @@ static int fw_get_filesystem_firmware(struct device *device,
> >>                 rc = kernel_read_file_from_path(path, &buf->data, &size,
> >>                                                 INT_MAX, READING_FIRMWARE);
> >>                 if (rc) {
> >> -                       dev_warn(device, "loading %s failed with error %d\n",
> >> +                       dev_dbg(device, "loading %s failed with error %d\n",
> >>                                  path, rc);
> >>                         continue;
> >>                 }
> >> --
> >> 2.7.0
> >>
> >
> > I think this should warn on non-ENOENT errors and dbg on ENOENT. What
> > do others think?
> 
> Agree, that looks better.

OK then this:

>From e63d19975787c0e237a47c17efd01e41b2a8e2fa Mon Sep 17 00:00:00 2001
From: "Luis R. Rodriguez" <mcgrof@...nel.org>
Date: Sat, 27 Feb 2016 14:58:08 -0800
Subject: [PATCH] firmware: change kernel read fail to dev_dbg()

When we now use the new kernel_read_file_from_path() we
are reporting a failure when we iterate over all the paths
possible for firmware. Before using kernel_read_file_from_path()
we only reported a failure once we confirmed a file existed
with filp_open() but failed with fw_read_file_contents().

With kernel_read_file_from_path() both are done for us and
we obviously are now reporting too much information given that
some optional paths will always fail and clutter the logs.

fw_get_filesystem_firmware() already has a check for failure
and uses an internal flag, FW_OPT_NO_WARN, but this does not
let us capture other unxpected errors. This enables that
as changed by Neil via commit:

"firmware: Be a bit more verbose about direct firmware loading failure"

Reported-by: Heiner Kallweit <hkallweit1@...il.com>
Cc: Neil Horman <nhorman@...driver.com>
Cc: Heiner Kallweit <hkallweit1@...il.com>
Cc: Mimi Zohar <zohar@...ux.vnet.ibm.com>
Cc: Kees Cook <keescook@...omium.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@...nel.org>
---
 drivers/base/firmware_class.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c
index 1cff832ab74e..9503a88b189b 100644
--- a/drivers/base/firmware_class.c
+++ b/drivers/base/firmware_class.c
@@ -328,8 +328,12 @@ static int fw_get_filesystem_firmware(struct device *device,
 		rc = kernel_read_file_from_path(path, &buf->data, &size,
 						INT_MAX, READING_FIRMWARE);
 		if (rc) {
-			dev_warn(device, "loading %s failed with error %d\n",
-				 path, rc);
+			if (rc == -ENOENT)
+				dev_dbg(device, "loading %s failed with error %d\n",
+					 path, rc);
+			else
+				dev_warn(device, "loading %s failed with error %d\n",
+					 path, rc);
 			continue;
 		}
 		dev_dbg(device, "direct-loading %s\n", buf->fw_id);
-- 
2.7.0