All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] ioctl: only memset non-NULL link settings
@ 2020-08-24  6:46 Hans-Christian Noren Egtvedt
  2020-08-27  8:23 ` Michal Kubecek
  0 siblings, 1 reply; 3+ messages in thread
From: Hans-Christian Noren Egtvedt @ 2020-08-24  6:46 UTC (permalink / raw)
  To: netdev; +Cc: Hans-Christian Noren Egtvedt

In commit bef780467fa ('ioctl: do not pass transceiver value back to
kernel') a regression slipped. If we have a kernel that does not support
the ETHTOOL_xLINKSETTINGS API, then the do_ioctl_glinksettings()
function will return a NULL pointer.

Hence before memset'ing the pointer to zero we must first check it is
valid, as NULL return is perfectly fine when running on old kernels.

Signed-off-by: Hans-Christian Noren Egtvedt <hegtvedt@cisco.com>
---
 ethtool.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/ethtool.c b/ethtool.c
index c4ad186..8267d6b 100644
--- a/ethtool.c
+++ b/ethtool.c
@@ -2908,8 +2908,10 @@ static int do_sset(struct cmd_context *ctx)
 		struct ethtool_link_usettings *link_usettings;
 
 		link_usettings = do_ioctl_glinksettings(ctx);
-		memset(&link_usettings->deprecated, 0,
-		       sizeof(link_usettings->deprecated));
+		if (link_usettings) {
+			memset(&link_usettings->deprecated, 0,
+			       sizeof(link_usettings->deprecated));
+		}
 		if (link_usettings == NULL)
 			link_usettings = do_ioctl_gset(ctx);
 		if (link_usettings == NULL) {
-- 
2.25.1


^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH] ioctl: only memset non-NULL link settings
  2020-08-24  6:46 [PATCH] ioctl: only memset non-NULL link settings Hans-Christian Noren Egtvedt
@ 2020-08-27  8:23 ` Michal Kubecek
  2020-08-27  9:50   ` Hans-Christian Egtvedt (hegtvedt)
  0 siblings, 1 reply; 3+ messages in thread
From: Michal Kubecek @ 2020-08-27  8:23 UTC (permalink / raw)
  To: Hans-Christian Noren Egtvedt; +Cc: netdev

[-- Attachment #1: Type: text/plain, Size: 1586 bytes --]

On Mon, Aug 24, 2020 at 08:46:30AM +0200, Hans-Christian Noren Egtvedt wrote:
> In commit bef780467fa ('ioctl: do not pass transceiver value back to
> kernel') a regression slipped. If we have a kernel that does not support
> the ETHTOOL_xLINKSETTINGS API, then the do_ioctl_glinksettings()
> function will return a NULL pointer.
> 
> Hence before memset'ing the pointer to zero we must first check it is
> valid, as NULL return is perfectly fine when running on old kernels.
> 

Please add a standard "Fixes:" line here, i.e.

Fixes: bef780467fa7 ("ioctl: do not pass transceiver value back to kernel")

> Signed-off-by: Hans-Christian Noren Egtvedt <hegtvedt@cisco.com>
> ---
>  ethtool.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/ethtool.c b/ethtool.c
> index c4ad186..8267d6b 100644
> --- a/ethtool.c
> +++ b/ethtool.c
> @@ -2908,8 +2908,10 @@ static int do_sset(struct cmd_context *ctx)
>  		struct ethtool_link_usettings *link_usettings;
>  
>  		link_usettings = do_ioctl_glinksettings(ctx);
> -		memset(&link_usettings->deprecated, 0,
> -		       sizeof(link_usettings->deprecated));
> +		if (link_usettings) {
> +			memset(&link_usettings->deprecated, 0,
> +			       sizeof(link_usettings->deprecated));
> +		}
>  		if (link_usettings == NULL)
>  			link_usettings = do_ioctl_gset(ctx);

This is correct fix but now we have two if statements with exactly
opposite conditions here; please combine them into one if-else.

Thank you,
Michal

>  		if (link_usettings == NULL) {
> -- 
> 2.25.1
> 

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] ioctl: only memset non-NULL link settings
  2020-08-27  8:23 ` Michal Kubecek
@ 2020-08-27  9:50   ` Hans-Christian Egtvedt (hegtvedt)
  0 siblings, 0 replies; 3+ messages in thread
From: Hans-Christian Egtvedt (hegtvedt) @ 2020-08-27  9:50 UTC (permalink / raw)
  To: Michal Kubecek; +Cc: netdev

On 27/08/2020 10:23, Michal Kubecek wrote:
> On Mon, Aug 24, 2020 at 08:46:30AM +0200, Hans-Christian Noren Egtvedt wrote:
>> In commit bef780467fa ('ioctl: do not pass transceiver value back to
>> kernel') a regression slipped. If we have a kernel that does not support
>> the ETHTOOL_xLINKSETTINGS API, then the do_ioctl_glinksettings()
>> function will return a NULL pointer.
>>
>> Hence before memset'ing the pointer to zero we must first check it is
>> valid, as NULL return is perfectly fine when running on old kernels.
>>
> 
> Please add a standard "Fixes:" line here, i.e.
> 
> Fixes: bef780467fa7 ("ioctl: do not pass transceiver value back to kernel")
> 
>> Signed-off-by: Hans-Christian Noren Egtvedt <hegtvedt@cisco.com>
>> ---
>>   ethtool.c | 6 ++++--
>>   1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/ethtool.c b/ethtool.c
>> index c4ad186..8267d6b 100644
>> --- a/ethtool.c
>> +++ b/ethtool.c
>> @@ -2908,8 +2908,10 @@ static int do_sset(struct cmd_context *ctx)
>>   		struct ethtool_link_usettings *link_usettings;
>>   
>>   		link_usettings = do_ioctl_glinksettings(ctx);
>> -		memset(&link_usettings->deprecated, 0,
>> -		       sizeof(link_usettings->deprecated));
>> +		if (link_usettings) {
>> +			memset(&link_usettings->deprecated, 0,
>> +			       sizeof(link_usettings->deprecated));
>> +		}
>>   		if (link_usettings == NULL)
>>   			link_usettings = do_ioctl_gset(ctx);
> 
> This is correct fix but now we have two if statements with exactly
> opposite conditions here; please combine them into one if-else.

Thank you for the review, a v2 will arrive shortly.

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2020-08-27  9:50 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-08-24  6:46 [PATCH] ioctl: only memset non-NULL link settings Hans-Christian Noren Egtvedt
2020-08-27  8:23 ` Michal Kubecek
2020-08-27  9:50   ` Hans-Christian Egtvedt (hegtvedt)

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.