* [PATCH net-next] iavf: Avoid a memory allocation in iavf_print_link_message() @ 2023-09-23 12:17 Christophe JAILLET 2023-09-25 8:11 ` Przemek Kitszel 2023-10-03 17:14 ` Jesse Brandeburg 0 siblings, 2 replies; 6+ messages in thread From: Christophe JAILLET @ 2023-09-23 12:17 UTC (permalink / raw) To: Jesse Brandeburg, Tony Nguyen, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni Cc: linux-kernel, kernel-janitors, Christophe JAILLET, intel-wired-lan, netdev IAVF_MAX_SPEED_STRLEN is only 13 and 'speed' is allocated and freed within iavf_print_link_message(). 'speed' is only used with some snprintf() and netdev_info() calls. So there is no real use to kzalloc()/free() it. Use the stack instead. This saves a memory allocation. Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr> --- drivers/net/ethernet/intel/iavf/iavf_virtchnl.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c index 8ce6389b5815..980dc69d7fbe 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c +++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c @@ -1389,18 +1389,14 @@ void iavf_disable_vlan_insertion_v2(struct iavf_adapter *adapter, u16 tpid) static void iavf_print_link_message(struct iavf_adapter *adapter) { struct net_device *netdev = adapter->netdev; + char speed[IAVF_MAX_SPEED_STRLEN]; int link_speed_mbps; - char *speed; if (!adapter->link_up) { netdev_info(netdev, "NIC Link is Down\n"); return; } - speed = kzalloc(IAVF_MAX_SPEED_STRLEN, GFP_KERNEL); - if (!speed) - return; - if (ADV_LINK_SUPPORT(adapter)) { link_speed_mbps = adapter->link_speed_mbps; goto print_link_msg; @@ -1452,7 +1448,6 @@ static void iavf_print_link_message(struct iavf_adapter *adapter) } netdev_info(netdev, "NIC Link is Up Speed is %s Full Duplex\n", speed); - kfree(speed); } /** -- 2.34.1 ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH net-next] iavf: Avoid a memory allocation in iavf_print_link_message() 2023-09-23 12:17 [PATCH net-next] iavf: Avoid a memory allocation in iavf_print_link_message() Christophe JAILLET @ 2023-09-25 8:11 ` Przemek Kitszel 2023-10-03 17:14 ` Jesse Brandeburg 1 sibling, 0 replies; 6+ messages in thread From: Przemek Kitszel @ 2023-09-25 8:11 UTC (permalink / raw) To: Christophe JAILLET, Jesse Brandeburg, Tony Nguyen, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni Cc: linux-kernel, kernel-janitors, intel-wired-lan, netdev On 9/23/23 14:17, Christophe JAILLET wrote: > IAVF_MAX_SPEED_STRLEN is only 13 and 'speed' is allocated and freed within > iavf_print_link_message(). > > 'speed' is only used with some snprintf() and netdev_info() calls. > > So there is no real use to kzalloc()/free() it. Use the stack instead. > This saves a memory allocation. > > Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr> > --- > drivers/net/ethernet/intel/iavf/iavf_virtchnl.c | 7 +------ > 1 file changed, 1 insertion(+), 6 deletions(-) > > diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c > index 8ce6389b5815..980dc69d7fbe 100644 > --- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c > +++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c > @@ -1389,18 +1389,14 @@ void iavf_disable_vlan_insertion_v2(struct iavf_adapter *adapter, u16 tpid) > static void iavf_print_link_message(struct iavf_adapter *adapter) > { > struct net_device *netdev = adapter->netdev; > + char speed[IAVF_MAX_SPEED_STRLEN]; > int link_speed_mbps; > - char *speed; > > if (!adapter->link_up) { > netdev_info(netdev, "NIC Link is Down\n"); > return; > } > > - speed = kzalloc(IAVF_MAX_SPEED_STRLEN, GFP_KERNEL); > - if (!speed) > - return; > - > if (ADV_LINK_SUPPORT(adapter)) { > link_speed_mbps = adapter->link_speed_mbps; > goto print_link_msg; > @@ -1452,7 +1448,6 @@ static void iavf_print_link_message(struct iavf_adapter *adapter) > } > > netdev_info(netdev, "NIC Link is Up Speed is %s Full Duplex\n", speed); > - kfree(speed); > } > > /** Looks fine, thanks! Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> I know that Jesse is fixing snprintf() calls currently, but I bet it's not conflicting. ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH net-next] iavf: Avoid a memory allocation in iavf_print_link_message() 2023-09-23 12:17 [PATCH net-next] iavf: Avoid a memory allocation in iavf_print_link_message() Christophe JAILLET 2023-09-25 8:11 ` Przemek Kitszel @ 2023-10-03 17:14 ` Jesse Brandeburg 2023-10-03 20:33 ` Christophe JAILLET 1 sibling, 1 reply; 6+ messages in thread From: Jesse Brandeburg @ 2023-10-03 17:14 UTC (permalink / raw) To: Christophe JAILLET, Tony Nguyen, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni Cc: linux-kernel, kernel-janitors, intel-wired-lan, netdev On 9/23/2023 5:17 AM, Christophe JAILLET wrote: > IAVF_MAX_SPEED_STRLEN is only 13 and 'speed' is allocated and freed within > iavf_print_link_message(). > > 'speed' is only used with some snprintf() and netdev_info() calls. > > So there is no real use to kzalloc()/free() it. Use the stack instead. > This saves a memory allocation. > > Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr> > --- > drivers/net/ethernet/intel/iavf/iavf_virtchnl.c | 7 +------ > 1 file changed, 1 insertion(+), 6 deletions(-) > > diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c > index 8ce6389b5815..980dc69d7fbe 100644 > --- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c > +++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c > @@ -1389,18 +1389,14 @@ void iavf_disable_vlan_insertion_v2(struct iavf_adapter *adapter, u16 tpid) > static void iavf_print_link_message(struct iavf_adapter *adapter) > { > struct net_device *netdev = adapter->netdev; > + char speed[IAVF_MAX_SPEED_STRLEN]; > int link_speed_mbps; > - char *speed; > > if (!adapter->link_up) { > netdev_info(netdev, "NIC Link is Down\n"); > return; > } > > - speed = kzalloc(IAVF_MAX_SPEED_STRLEN, GFP_KERNEL); > - if (!speed) > - return; > - > if (ADV_LINK_SUPPORT(adapter)) { > link_speed_mbps = adapter->link_speed_mbps; > goto print_link_msg; > @@ -1452,7 +1448,6 @@ static void iavf_print_link_message(struct iavf_adapter *adapter) > } > > netdev_info(netdev, "NIC Link is Up Speed is %s Full Duplex\n", speed); > - kfree(speed); > } > > /** Hi Christophe! I had a slightly different proposal that gets rid of all the -Wformat=2 warnings in this code by using kasprintf to handle the varying string lengths. any thoughts about this instead and drop yours? I'm less worried about the "extra allocation" here in this function since it's slow path, and the same comment applies to your patch as well. your patch still shows these errors > drivers/net/ethernet/intel/iavf/iavf_virtchnl.c: In function ‘iavf_virtchnl_completion’: > drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1446:60: warning: ‘%s’ directive output may be truncated writing 4 bytes into a region of size between 1 and 11 [-Wformat-truncation=] > 1446 | snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%d %s", > | ^~ > 1447 | link_speed_mbps, "Mbps"); > | ~~~~~~ > In function ‘iavf_print_link_message’, > inlined from ‘iavf_virtchnl_completion’ at drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1965:4: > drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1446:17: note: ‘snprintf’ output between 7 and 17 bytes into a destination of size 13 > 1446 | snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%d %s", > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > 1447 | link_speed_mbps, "Mbps"); > | ~~~~~~~~~~~~~~~~~~~~~~~~ <my iavf patch pasted as a quote so my mail client won't wrap the lines...> > diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c > index 8ce6389b5815..82b84a93bcc8 100644 > --- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c > +++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c > @@ -1378,8 +1378,6 @@ void iavf_disable_vlan_insertion_v2(struct iavf_adapter *adapter, u16 tpid) > VIRTCHNL_OP_DISABLE_VLAN_INSERTION_V2); > } > > -#define IAVF_MAX_SPEED_STRLEN 13 > - > /** > * iavf_print_link_message - print link up or down > * @adapter: adapter structure > @@ -1397,10 +1395,6 @@ static void iavf_print_link_message(struct iavf_adapter *adapter) > return; > } > > - speed = kzalloc(IAVF_MAX_SPEED_STRLEN, GFP_KERNEL); > - if (!speed) > - return; > - > if (ADV_LINK_SUPPORT(adapter)) { > link_speed_mbps = adapter->link_speed_mbps; > goto print_link_msg; > @@ -1438,17 +1432,17 @@ static void iavf_print_link_message(struct iavf_adapter *adapter) > > print_link_msg: > if (link_speed_mbps > SPEED_1000) { > - if (link_speed_mbps == SPEED_2500) > - snprintf(speed, IAVF_MAX_SPEED_STRLEN, "2.5 Gbps"); > - else > + if (link_speed_mbps == SPEED_2500) { > + speed = kasprintf(GFP_KERNEL, "%s", "2.5 Gbps"); > + } else { > /* convert to Gbps inline */ > - snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%d %s", > - link_speed_mbps / 1000, "Gbps"); > + speed = kasprintf(GFP_KERNEL, "%d Gbps", > + link_speed_mbps / 1000); > + } > } else if (link_speed_mbps == SPEED_UNKNOWN) { > - snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%s", "Unknown Mbps"); > + speed = kasprintf(GFP_KERNEL, "%s", "Unknown Mbps"); > } else { > - snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%d %s", > - link_speed_mbps, "Mbps"); > + speed = kasprintf(GFP_KERNEL, "%d Mbps", link_speed_mbps); > } > > netdev_info(netdev, "NIC Link is Up Speed is %s Full Duplex\n", speed); ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH net-next] iavf: Avoid a memory allocation in iavf_print_link_message() 2023-10-03 17:14 ` Jesse Brandeburg @ 2023-10-03 20:33 ` Christophe JAILLET 2023-10-03 23:01 ` Jesse Brandeburg 0 siblings, 1 reply; 6+ messages in thread From: Christophe JAILLET @ 2023-10-03 20:33 UTC (permalink / raw) To: Jesse Brandeburg, Tony Nguyen, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni Cc: linux-kernel, kernel-janitors, intel-wired-lan, netdev Le 03/10/2023 à 19:14, Jesse Brandeburg a écrit : > On 9/23/2023 5:17 AM, Christophe JAILLET wrote: >> IAVF_MAX_SPEED_STRLEN is only 13 and 'speed' is allocated and freed within >> iavf_print_link_message(). >> >> 'speed' is only used with some snprintf() and netdev_info() calls. >> >> So there is no real use to kzalloc()/free() it. Use the stack instead. >> This saves a memory allocation. >> >> Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr> >> --- >> drivers/net/ethernet/intel/iavf/iavf_virtchnl.c | 7 +------ >> 1 file changed, 1 insertion(+), 6 deletions(-) >> >> diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c >> index 8ce6389b5815..980dc69d7fbe 100644 >> --- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c >> +++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c >> @@ -1389,18 +1389,14 @@ void iavf_disable_vlan_insertion_v2(struct iavf_adapter *adapter, u16 tpid) >> static void iavf_print_link_message(struct iavf_adapter *adapter) >> { >> struct net_device *netdev = adapter->netdev; >> + char speed[IAVF_MAX_SPEED_STRLEN]; >> int link_speed_mbps; >> - char *speed; >> >> if (!adapter->link_up) { >> netdev_info(netdev, "NIC Link is Down\n"); >> return; >> } >> >> - speed = kzalloc(IAVF_MAX_SPEED_STRLEN, GFP_KERNEL); >> - if (!speed) >> - return; >> - >> if (ADV_LINK_SUPPORT(adapter)) { >> link_speed_mbps = adapter->link_speed_mbps; >> goto print_link_msg; >> @@ -1452,7 +1448,6 @@ static void iavf_print_link_message(struct iavf_adapter *adapter) >> } >> >> netdev_info(netdev, "NIC Link is Up Speed is %s Full Duplex\n", speed); >> - kfree(speed); >> } >> >> /** > > Hi Christophe! > > I had a slightly different proposal that gets rid of all the -Wformat=2 > warnings in this code by using kasprintf to handle the varying string > lengths. > > any thoughts about this instead and drop yours? I'm less worried about > the "extra allocation" here in this function since it's slow path, and > the same comment applies to your patch as well. kasprintf() is much better. > > your patch still shows these errors I built-tested the patch before sending, so this is strange. However, I got a similar feedback from Greg KH and the "kernel test robot" for another similar patch. What version of gcc do you use? I use 12.3.0, and I suspect that the value range algorithm or how the diagnostic is done has been improved in recent gcc. The other report was from 11.3.0. CJ >> drivers/net/ethernet/intel/iavf/iavf_virtchnl.c: In function ‘iavf_virtchnl_completion’: >> drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1446:60: warning: ‘%s’ directive output may be truncated writing 4 bytes into a region of size between 1 and 11 [-Wformat-truncation=] >> 1446 | snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%d %s", >> | ^~ >> 1447 | link_speed_mbps, "Mbps"); >> | ~~~~~~ >> In function ‘iavf_print_link_message’, >> inlined from ‘iavf_virtchnl_completion’ at drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1965:4: >> drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1446:17: note: ‘snprintf’ output between 7 and 17 bytes into a destination of size 13 >> 1446 | snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%d %s", >> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >> 1447 | link_speed_mbps, "Mbps"); >> | ~~~~~~~~~~~~~~~~~~~~~~~~ > > > <my iavf patch pasted as a quote so my mail client won't wrap the lines...> > > >> diff --git a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c >> index 8ce6389b5815..82b84a93bcc8 100644 >> --- a/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c >> +++ b/drivers/net/ethernet/intel/iavf/iavf_virtchnl.c >> @@ -1378,8 +1378,6 @@ void iavf_disable_vlan_insertion_v2(struct iavf_adapter *adapter, u16 tpid) >> VIRTCHNL_OP_DISABLE_VLAN_INSERTION_V2); >> } >> >> -#define IAVF_MAX_SPEED_STRLEN 13 >> - >> /** >> * iavf_print_link_message - print link up or down >> * @adapter: adapter structure >> @@ -1397,10 +1395,6 @@ static void iavf_print_link_message(struct iavf_adapter *adapter) >> return; >> } >> >> - speed = kzalloc(IAVF_MAX_SPEED_STRLEN, GFP_KERNEL); >> - if (!speed) >> - return; >> - >> if (ADV_LINK_SUPPORT(adapter)) { >> link_speed_mbps = adapter->link_speed_mbps; >> goto print_link_msg; >> @@ -1438,17 +1432,17 @@ static void iavf_print_link_message(struct iavf_adapter *adapter) >> >> print_link_msg: >> if (link_speed_mbps > SPEED_1000) { >> - if (link_speed_mbps == SPEED_2500) >> - snprintf(speed, IAVF_MAX_SPEED_STRLEN, "2.5 Gbps"); >> - else >> + if (link_speed_mbps == SPEED_2500) { >> + speed = kasprintf(GFP_KERNEL, "%s", "2.5 Gbps"); >> + } else { >> /* convert to Gbps inline */ >> - snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%d %s", >> - link_speed_mbps / 1000, "Gbps"); >> + speed = kasprintf(GFP_KERNEL, "%d Gbps", >> + link_speed_mbps / 1000); >> + } >> } else if (link_speed_mbps == SPEED_UNKNOWN) { >> - snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%s", "Unknown Mbps"); >> + speed = kasprintf(GFP_KERNEL, "%s", "Unknown Mbps"); >> } else { >> - snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%d %s", >> - link_speed_mbps, "Mbps"); >> + speed = kasprintf(GFP_KERNEL, "%d Mbps", link_speed_mbps); >> } >> >> netdev_info(netdev, "NIC Link is Up Speed is %s Full Duplex\n", speed); > > > ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH net-next] iavf: Avoid a memory allocation in iavf_print_link_message() 2023-10-03 20:33 ` Christophe JAILLET @ 2023-10-03 23:01 ` Jesse Brandeburg 2023-10-04 8:04 ` Dan Carpenter 0 siblings, 1 reply; 6+ messages in thread From: Jesse Brandeburg @ 2023-10-03 23:01 UTC (permalink / raw) To: Christophe JAILLET, Tony Nguyen, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni Cc: linux-kernel, kernel-janitors, intel-wired-lan, netdev On 10/3/2023 1:33 PM, Christophe JAILLET wrote: > kasprintf() is much better. cool! I just sent the patches and cc'd you earlier today. >> >> your patch still shows these errors > > I built-tested the patch before sending, so this is strange. > > However, I got a similar feedback from Greg KH and the "kernel test > robot" for another similar patch. > > What version of gcc do you use? > I use 12.3.0, and I suspect that the value range algorithm or how the > diagnostic is done has been improved in recent gcc. Fedora gcc 12.3.1, with W=1 flag gcc version 12.3.1 20230508 (Red Hat 12.3.1-1) (GCC) [linux]$ make W=1 M=drivers/net/ethernet/intel/iavf CC [M] drivers/net/ethernet/intel/iavf/iavf_main.o CC [M] drivers/net/ethernet/intel/iavf/iavf_ethtool.o CC [M] drivers/net/ethernet/intel/iavf/iavf_virtchnl.o CC [M] drivers/net/ethernet/intel/iavf/iavf_fdir.o CC [M] drivers/net/ethernet/intel/iavf/iavf_adv_rss.o CC [M] drivers/net/ethernet/intel/iavf/iavf_txrx.o CC [M] drivers/net/ethernet/intel/iavf/iavf_common.o CC [M] drivers/net/ethernet/intel/iavf/iavf_adminq.o CC [M] drivers/net/ethernet/intel/iavf/iavf_client.o drivers/net/ethernet/intel/iavf/iavf_virtchnl.c: In function ‘iavf_virtchnl_completion’: drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1446:60: warning: ‘%s’ directive output may be truncated writing 4 bytes into a region of size between 1 and 11 [-Wformat-truncation=] 1446 | snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%d %s", | ^~ 1447 | link_speed_mbps, "Mbps"); | ~~~~~~ In function ‘iavf_print_link_message’, inlined from ‘iavf_virtchnl_completion’ at drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1965:4: drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1446:17: note: ‘snprintf’ output between 7 and 17 bytes into a destination of size 13 1446 | snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%d %s", | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1447 | link_speed_mbps, "Mbps"); | ~~~~~~~~~~~~~~~~~~~~~~~~ ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH net-next] iavf: Avoid a memory allocation in iavf_print_link_message() 2023-10-03 23:01 ` Jesse Brandeburg @ 2023-10-04 8:04 ` Dan Carpenter 0 siblings, 0 replies; 6+ messages in thread From: Dan Carpenter @ 2023-10-04 8:04 UTC (permalink / raw) To: Jesse Brandeburg Cc: Christophe JAILLET, Tony Nguyen, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni, linux-kernel, kernel-janitors, intel-wired-lan, netdev On Tue, Oct 03, 2023 at 04:01:18PM -0700, Jesse Brandeburg wrote: > On 10/3/2023 1:33 PM, Christophe JAILLET wrote: > > kasprintf() is much better. > > cool! I just sent the patches and cc'd you earlier today. > > > > > > > your patch still shows these errors > > > > I built-tested the patch before sending, so this is strange. > > > > However, I got a similar feedback from Greg KH and the "kernel test > > robot" for another similar patch. > > > > What version of gcc do you use? > > I use 12.3.0, and I suspect that the value range algorithm or how the > > diagnostic is done has been improved in recent gcc. > > Fedora gcc 12.3.1, with W=1 flag > > gcc version 12.3.1 20230508 (Red Hat 12.3.1-1) (GCC) > > [linux]$ make W=1 M=drivers/net/ethernet/intel/iavf > CC [M] drivers/net/ethernet/intel/iavf/iavf_main.o > CC [M] drivers/net/ethernet/intel/iavf/iavf_ethtool.o > CC [M] drivers/net/ethernet/intel/iavf/iavf_virtchnl.o > CC [M] drivers/net/ethernet/intel/iavf/iavf_fdir.o > CC [M] drivers/net/ethernet/intel/iavf/iavf_adv_rss.o > CC [M] drivers/net/ethernet/intel/iavf/iavf_txrx.o > CC [M] drivers/net/ethernet/intel/iavf/iavf_common.o > CC [M] drivers/net/ethernet/intel/iavf/iavf_adminq.o > CC [M] drivers/net/ethernet/intel/iavf/iavf_client.o > drivers/net/ethernet/intel/iavf/iavf_virtchnl.c: In function > ‘iavf_virtchnl_completion’: > drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1446:60: warning: ‘%s’ > directive output may be truncated writing 4 bytes into a region of size > between 1 and 11 [-Wformat-truncation=] > 1446 | snprintf(speed, IAVF_MAX_SPEED_STRLEN, "%d %s", > | ^~ > 1447 | link_speed_mbps, "Mbps"); > | ~~~~~~ GCC is kind of crap at static analysis, right? Smatch would know that this at most 11 characters long. It's kind of laziness for GCC to print this warning. If you complained to me about a false positive like this in Smatch I would at least think about various ways to silence it. But I probably wouldn't write a check for this anyway because I don't view truncating strings as a note worthy bug... Smatch also gets stuff wrong, but in that case I just always encourage people to mark the warning as old news and move on. Only new warnings are interesting. I feel like as we incorporate more and more static analysis into our processes we're going to have to give up on trying to keep every static checker happy. regards, dan carpenter ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2023-10-04 8:04 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2023-09-23 12:17 [PATCH net-next] iavf: Avoid a memory allocation in iavf_print_link_message() Christophe JAILLET 2023-09-25 8:11 ` Przemek Kitszel 2023-10-03 17:14 ` Jesse Brandeburg 2023-10-03 20:33 ` Christophe JAILLET 2023-10-03 23:01 ` Jesse Brandeburg 2023-10-04 8:04 ` Dan Carpenter
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®