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

Groups > linux.kernel > #1218446 > unrolled thread

Re: [PATCHv2 18/19] staging: rtl8192u: r8192U_core: fix use ether_addr_copy() over memcpy() code style issue

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCHv2 18/19] staging: rtl8192u: r8192U_core: fix use  ether_addr_copy() over memcpy() code style issue Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2015-09-03 19:10 +0200
    Re: [PATCHv2 18/19] staging: rtl8192u: r8192U_core: fix use  ether_addr_copy() over memcpy() code style issue Raphaël Beamonte <raphael.beamonte@gmail.com> - 2015-09-06 00:20 +0200
      Re: [PATCHv2 18/19] staging: rtl8192u: r8192U_core: fix use  ether_addr_copy() over memcpy() code style issue Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2015-09-06 02:50 +0200
        [PATCHv2] staging: rtl8192u: r8192U_core: fix use ether_addr_copy() over memcpy() code style issue Raphaël Beamonte <raphael.beamonte@gmail.com> - 2015-09-06 04:20 +0200

#1218446 — Re: [PATCHv2 18/19] staging: rtl8192u: r8192U_core: fix use ether_addr_copy() over memcpy() code style issue

On Tue, Aug 18, 2015 at 12:58:19PM -0400, Raphaël Beamonte wrote:
> Prefer ether_addr_copy() over memcpy() if the Ethernet addresses are __aligned(2)

You need to prove that they are aligned before I can take this patch :(
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [next] | [standalone]

#1219652

2015-09-03 13:00 GMT-04:00 Greg Kroah-Hartman <gregkh@linuxfoundation.org>:
> You need to prove that they are aligned before I can take this patch :(

Right!

I looked at those structs using pahole. The structs used are
net_device from include/linux/netdevice.h and sockaddr from
include/linux/socket.h, in which we use the elements dev_addr and
sa_data respectively.

Using pahole, the output for both those elements in the structs are:
unsigned char *            dev_addr;             /*   888     8 */
char                       sa_data[14];          /*     2    14 */

If I understand correctly how pahole is outputing the information and
what a u16 alignment means in the struct, I should, for both those
cases, have a multiple of 2 (sizeof(u16)) as first value in the
comment following the line (first value being the position of the
element in the struct, second value its size). So if I understood
correctly, in both those cases, it's working: they're u16 aligned.

Also, to be sure, I searched in the kernel sources and found multiple
files that are currently using both those structs and ether_addr_copy:
(line numbers are those for commit 34fca4b of your staging-testing
tree)
drivers/net/ethernet/broadcom/genet/bcmgenet.c (line 2957)
drivers/net/ethernet/brocade/bna/bnad.c (line 3250)
drivers/net/ethernet/emulex/benet/be_main.c (line 315)
drivers/net/ethernet/ezchip/nps_enet.c (line 421)
drivers/net/ethernet/ibm/ibmveth.c (line 1460)
drivers/net/ethernet/intel/fm10k/fm10k_netdev.c (line 911)
drivers/net/ethernet/intel/i40e/i40e_main.c (line 1518)
drivers/net/ethernet/mellanox/mlx5/core/en_main.c (line 1796)
drivers/net/usb/lan78xx.c (line 1777)
drivers/staging/rtl8192e/rtl8192e/rtl_core.c (line 2285)
drivers/staging/rtl8712/os_intfs.c (line 181)
net/8021q/vlan_dev.c (line 338)
net/batman-adv/soft-interface.c (line 140)
net/dsa/slave.c (line 198)

Given that information, it should be fine for that change in rtl8192u!
But if I'm mistaken somewhere please tell me!
Should I update my patch with those information in the changelog? If
so, how verbose do you expect it to be?

Thanks!
-R
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1219665

On Sat, Sep 05, 2015 at 06:16:16PM -0400, Raphaël Beamonte wrote:
> 2015-09-03 13:00 GMT-04:00 Greg Kroah-Hartman <gregkh@linuxfoundation.org>:
> > You need to prove that they are aligned before I can take this patch :(
> 
> Right!
> 
> I looked at those structs using pahole. The structs used are
> net_device from include/linux/netdevice.h and sockaddr from
> include/linux/socket.h, in which we use the elements dev_addr and
> sa_data respectively.
> 
> Using pahole, the output for both those elements in the structs are:
> unsigned char *            dev_addr;             /*   888     8 */
> char                       sa_data[14];          /*     2    14 */
> 
> If I understand correctly how pahole is outputing the information and
> what a u16 alignment means in the struct, I should, for both those
> cases, have a multiple of 2 (sizeof(u16)) as first value in the
> comment following the line (first value being the position of the
> element in the struct, second value its size). So if I understood
> correctly, in both those cases, it's working: they're u16 aligned.
> 
> Also, to be sure, I searched in the kernel sources and found multiple
> files that are currently using both those structs and ether_addr_copy:
> (line numbers are those for commit 34fca4b of your staging-testing
> tree)
> drivers/net/ethernet/broadcom/genet/bcmgenet.c (line 2957)
> drivers/net/ethernet/brocade/bna/bnad.c (line 3250)
> drivers/net/ethernet/emulex/benet/be_main.c (line 315)
> drivers/net/ethernet/ezchip/nps_enet.c (line 421)
> drivers/net/ethernet/ibm/ibmveth.c (line 1460)
> drivers/net/ethernet/intel/fm10k/fm10k_netdev.c (line 911)
> drivers/net/ethernet/intel/i40e/i40e_main.c (line 1518)
> drivers/net/ethernet/mellanox/mlx5/core/en_main.c (line 1796)
> drivers/net/usb/lan78xx.c (line 1777)
> drivers/staging/rtl8192e/rtl8192e/rtl_core.c (line 2285)
> drivers/staging/rtl8712/os_intfs.c (line 181)
> net/8021q/vlan_dev.c (line 338)
> net/batman-adv/soft-interface.c (line 140)
> net/dsa/slave.c (line 198)
> 
> Given that information, it should be fine for that change in rtl8192u!
> But if I'm mistaken somewhere please tell me!
> Should I update my patch with those information in the changelog?

Yes.

> If so, how verbose do you expect it to be?

Very :)

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1219684 — [PATCHv2] staging: rtl8192u: r8192U_core: fix use ether_addr_copy() over memcpy() code style issue

Prefer ether_addr_copy() over memcpy() if the Ethernet addresses
are __aligned(2)

The values used are stored as dev_addr in net_device (declared in
include/linux/netdevice.h) and sa_data in sockaddr (declared in
include/linux/socket.h). Both these elements are u16 aligned as
shown by using pahole (position must be a multiple of sizeof(u16)):
unsigned char *            dev_addr;             /*   888     8 */
char                       sa_data[14];          /*     2    14 */

It is thus safe to use ether_addr_copy() instead of memcpy() for
that call, as it is already done in multiple files in the Linux
kernel sources:
  drivers/net/ethernet/broadcom/genet/bcmgenet.c
  drivers/net/ethernet/brocade/bna/bnad.c
  drivers/net/ethernet/emulex/benet/be_main.c
  drivers/net/ethernet/ezchip/nps_enet.c
  drivers/net/ethernet/ibm/ibmveth.c
  drivers/net/ethernet/intel/fm10k/fm10k_netdev.c
  drivers/net/ethernet/intel/i40e/i40e_main.c
  drivers/net/ethernet/mellanox/mlx5/core/en_main.c
  drivers/net/usb/lan78xx.c
  net/8021q/vlan_dev.c
  net/batman-adv/soft-interface.c
  net/dsa/slave.c

Signed-off-by: Raphaël Beamonte <raphael.beamonte@gmail.com>
---
 drivers/staging/rtl8192u/r8192U_core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/staging/rtl8192u/r8192U_core.c b/drivers/staging/rtl8192u/r8192U_core.c
index 76255cb..b143b36 100644
--- a/drivers/staging/rtl8192u/r8192U_core.c
+++ b/drivers/staging/rtl8192u/r8192U_core.c
@@ -3469,7 +3469,7 @@ static int r8192_set_mac_adr(struct net_device *dev, void *mac)
 
 	down(&priv->wx_sem);
 
-	memcpy(dev->dev_addr, addr->sa_data, ETH_ALEN);
+	ether_addr_copy(dev->dev_addr, addr->sa_data);
 
 	schedule_work(&priv->reset_wq);
 	up(&priv->wx_sem);
-- 
2.5.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web