Re: [RTnet-users] r8169 / rtl8168/ 8111 driver
On 11/10/2012 09:24, Jan Kiszka wrote: > On 2012-10-11 08:40, Stéphane ANCELOT wrote: >> Hi, >> >> This driver has been derived from the 2.6.34 vanilla kernel. >> >> 2 years ago, there was a lack with 8168 support in existing rt_r8169 >> driver, from our viewpoint there was more added value porting the >> vanilla kernel one, and if I remember some caveats with MSI interrupts. >> >> I think you should not replace the existing rt_r8169 driver, but use it >> as an additional driver for 8111/8168 , it should also drive r8169 , but >> I have not tried it. >> >> I am running this driver for more than 2 years, without any problems >> (but I agree, only one target platform). >> >> The improvements needed are regarding driver's behaviour in a bad >> network context (equipment failure, connection problems, and so on), >> but this is a general feature and is not only related to this driver, >> and may depend on your application. > I'm willing to merge patches to fix the existing driver or replace it > with something better, but no longer carry drivers with overlapping > support. The E1000 mess was already painful enough (still need to decide > on the E1000-new). > > Looking at the diff and the age of the rt_r8169, reviewing, fixing and > testing your driver may be the better option. > However, from a first glance: > - rtdm_irq_disable/enable is fishy, can you explain its need? I had not read completely the rtdm_irq_enable manual...looking at it, yes, I agree it sounds better masking IT in registers. > - rtl8169_tx_timeout -> schedule_work is called from RTDM IRQ context In my case, this was not a concern, since we disabled FifoOverflow IT. > - rtl8169_start_xmit leaks a lock on error you're right. > So there is some work remaining to make it ready for upstream. As I have > test hardware around, I can help with validating the 8169 case. > > Jan > > -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev ___ RTnet-users mailing list RTnet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rtnet-users
Re: [RTnet-users] r8169 / rtl8168/ 8111 driver
On 2012-10-11 08:40, Stéphane ANCELOT wrote: > Hi, > > This driver has been derived from the 2.6.34 vanilla kernel. > > 2 years ago, there was a lack with 8168 support in existing rt_r8169 > driver, from our viewpoint there was more added value porting the > vanilla kernel one, and if I remember some caveats with MSI interrupts. > > I think you should not replace the existing rt_r8169 driver, but use it > as an additional driver for 8111/8168 , it should also drive r8169 , but > I have not tried it. > > I am running this driver for more than 2 years, without any problems > (but I agree, only one target platform). > > The improvements needed are regarding driver's behaviour in a bad > network context (equipment failure, connection problems, and so on), > but this is a general feature and is not only related to this driver, > and may depend on your application. I'm willing to merge patches to fix the existing driver or replace it with something better, but no longer carry drivers with overlapping support. The E1000 mess was already painful enough (still need to decide on the E1000-new). Looking at the diff and the age of the rt_r8169, reviewing, fixing and testing your driver may be the better option. However, from a first glance: - rtdm_irq_disable/enable is fishy, can you explain its need? - rtl8169_tx_timeout -> schedule_work is called from RTDM IRQ context - rtl8169_start_xmit leaks a lock on error So there is some work remaining to make it ready for upstream. As I have test hardware around, I can help with validating the 8169 case. Jan signature.asc Description: OpenPGP digital signature -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev___ RTnet-users mailing list RTnet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rtnet-users
Re: [RTnet-users] r8169 / rtl8168/ 8111 driver
Hi, This driver has been derived from the 2.6.34 vanilla kernel. 2 years ago, there was a lack with 8168 support in existing rt_r8169 driver, from our viewpoint there was more added value porting the vanilla kernel one, and if I remember some caveats with MSI interrupts. I think you should not replace the existing rt_r8169 driver, but use it as an additional driver for 8111/8168 , it should also drive r8169 , but I have not tried it. I am running this driver for more than 2 years, without any problems (but I agree, only one target platform). The improvements needed are regarding driver's behaviour in a bad network context (equipment failure, connection problems, and so on), but this is a general feature and is not only related to this driver, and may depend on your application. Regards, S.Ancelot On 11/10/2012 06:57, Jan Kiszka wrote: > On 2012-10-10 14:28, Stéphane ANCELOT wrote: >> Hi everybody ! >> >> You can use the following RT net driver for r8169 / 8168 / 8111 >> > Thanks for sharing, but how does this relate to the existing rt_r8169? > Can't the latter be augmented, if required, to fulfill your needs? Would > be preferred. > > Jan > >> I know it is working fine for r8111/8168 component we are using. >> >> The file is enclosed. >> >> Regards, >> S.Ancelot -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev ___ RTnet-users mailing list RTnet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rtnet-users
Re: [RTnet-users] r8169 / rtl8168/ 8111 driver
On 2012-10-10 14:28, Stéphane ANCELOT wrote: > Hi everybody ! > > You can use the following RT net driver for r8169 / 8168 / 8111 > Thanks for sharing, but how does this relate to the existing rt_r8169? Can't the latter be augmented, if required, to fulfill your needs? Would be preferred. Jan > > I know it is working fine for r8111/8168 component we are using. > > The file is enclosed. > > Regards, > S.Ancelot signature.asc Description: OpenPGP digital signature -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev___ RTnet-users mailing list RTnet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rtnet-users