Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1675899 > unrolled thread

[PATCH] iwlwifi: mvm: fix iwl_mvm_sar_find_wifi_pkg corner case

Started byArnd Bergmann <arnd@arndb.de>
First post2017-06-27 17:30 +0200
Last post2017-06-28 14:10 +0200
Articles 2 — 2 participants

Back to article view | Back to linux.kernel


Contents

  [PATCH] iwlwifi: mvm: fix iwl_mvm_sar_find_wifi_pkg corner case Arnd Bergmann <arnd@arndb.de> - 2017-06-27 17:30 +0200
    Re: [PATCH] iwlwifi: mvm: fix iwl_mvm_sar_find_wifi_pkg corner case Luca Coelho <luca@coelho.fi> - 2017-06-28 14:10 +0200

#1675899 — [PATCH] iwlwifi: mvm: fix iwl_mvm_sar_find_wifi_pkg corner case

FromArnd Bergmann <arnd@arndb.de>
Date2017-06-27 17:30 +0200
Subject[PATCH] iwlwifi: mvm: fix iwl_mvm_sar_find_wifi_pkg corner case
Message-ID<tX0Bd-2nz-43@gated-at.bofh.it>
gcc warns about what it thinks is an uninitialized variable
access:

drivers/net/wireless/intel/iwlwifi/mvm/fw.c: In function 'iwl_mvm_sar_find_wifi_pkg.isra.14':
drivers/net/wireless/intel/iwlwifi/mvm/fw.c:1102:5: error: 'wifi_pkg' may be used uninitialized in this function [-Werror=maybe-uninitialized]

That problem cannot really happen, as we check data->package.count
to ensure that the loop is entered at least once.
However, something that can indeed happen is returning an incorrect
wifi_pkg pointer in case none of the elements are what we are looking
for.

This modifies the loop again, to only return a correct object, and
to shut up that warning.

Fixes: c386dacb4ed6 ("iwlwifi: mvm: refactor SAR init to prepare for dynamic SAR")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
 drivers/net/wireless/intel/iwlwifi/mvm/fw.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/net/wireless/intel/iwlwifi/mvm/fw.c b/drivers/net/wireless/intel/iwlwifi/mvm/fw.c
index 24cc406d87ef..730c7f68c0b3 100644
--- a/drivers/net/wireless/intel/iwlwifi/mvm/fw.c
+++ b/drivers/net/wireless/intel/iwlwifi/mvm/fw.c
@@ -1094,14 +1094,12 @@ static union acpi_object *iwl_mvm_sar_find_wifi_pkg(struct iwl_mvm *mvm,
 		domain = &wifi_pkg->package.elements[0];
 		if (domain->type == ACPI_TYPE_INTEGER &&
 		    domain->integer.value == ACPI_WIFI_DOMAIN)
-			break;
-
-		wifi_pkg = NULL;
+			goto found;
 	}
 
-	if (!wifi_pkg)
-		return ERR_PTR(-ENOENT);
+	return ERR_PTR(-ENOENT);
 
+found:
 	return wifi_pkg;
 }
 
-- 
2.9.0

[toc] | [next] | [standalone]


#1676606

FromLuca Coelho <luca@coelho.fi>
Date2017-06-28 14:10 +0200
Message-ID<tXjXd-6DX-33@gated-at.bofh.it>
In reply to#1675899
On Tue, 2017-06-27 at 17:24 +0200, Arnd Bergmann wrote:
> gcc warns about what it thinks is an uninitialized variable
> access:
> 
> drivers/net/wireless/intel/iwlwifi/mvm/fw.c: In function 'iwl_mvm_sar_find_wifi_pkg.isra.14':
> drivers/net/wireless/intel/iwlwifi/mvm/fw.c:1102:5: error: 'wifi_pkg' may be used uninitialized in this function [-Werror=maybe-uninitialized]
> 
> That problem cannot really happen, as we check data->package.count
> to ensure that the loop is entered at least once.
> However, something that can indeed happen is returning an incorrect
> wifi_pkg pointer in case none of the elements are what we are looking
> for.
> 
> This modifies the loop again, to only return a correct object, and
> to shut up that warning.
> 
> Fixes: c386dacb4ed6 ("iwlwifi: mvm: refactor SAR init to prepare for dynamic SAR")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---

Thanks, Arnd!

I've pushed this to our internal tree and it will eventually reach the
mainline, via our normal upstreaming process.


--
Cheers,
Luca.

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web