Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Hi Patrick, Julien Since linux-image-3.1.0 arrived in unstable (many thanks Ben), perhaps it would be nice if you could try it with your NIC so we learn how well new r8169 work with your card. Could you please provide some feedback? Because r8168 qualifies for non-free I would like to have it there only if none of r8169 works. If working alternative available in newer kernels, we simply don't need r8168 and this will suit Ben. Thanks for http://code.google.com/p/r8168/ which seems to be right upstream home page which I wasn't aware of. I don't have problematic NIC so just like Andreas I can only test the package but not the driver. Regards, Dmitry. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/canbdodvopojr1ee9j+sgeuwvb6+u4nnyxgdupik4oevzt77...@mail.gmail.com
Re: Bug#642198: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
On 2011-11-13 16:59, Patrick Matthäi wrote: > Am 13.11.2011 15:08, schrieb Andreas Beckmann: >> I just revived my old ITP http://bugs.debian.org/642198 Updated packaging is available at git://github.com/anbe42/r8168.git Patrick, you may want to test it. Also testing to switch back from r8168 to r8169 would be interesting. > But btw isn't http://code.google.com/p/r8168/ the better homepage for it? That's an unofficial mirror, so I wouldn't take it for the Homepage. But at least we can use it for the watch file. > [...] and I am still available to sponsor the upload :) I still offer to maintain the package (or co-maintain it with Dmitry). But I don't have any NIC that requires r8168 (on a quick glance I didn't even find anything using r8169), so I can't test beyond module compilation. If you want to see it rather in non-free than in main, please say so. Andreas -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4ec10d96.4040...@abeckmann.de
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Am 13.11.2011 15:08, schrieb Andreas Beckmann: On 2011-11-12 19:04, Patrick Matthäi wrote: @Andreas and Dmitry: You may cooperate on packaging or decide, who wants to maintain it in the future. I just revived my old ITP http://bugs.debian.org/642198 and tried to put my things in a git repository: Only git://git.debian.org/~anbe-guest/r8168.git seems to work so far. @Ben: I still think this driver should be added as alternative driver to the archive, until r8169 will do its job for the promoted PCIIDs correctly. One possible solution I see would be to upload r8168 to sid (or experimental) and add a severity=serious bug "r8168 should not enter testing - let's fix r8169 in the kernel instead" to prevent migration. Documentation should include detailed instructions how to properly report bugs for the in-kernel r8169 if it does not work (but r8168 does). I do not agree. The package description and README should state, that it is an alternate driver which should be used, if r8169 does not work. It SHOULD enter wheezy as alternative, if r8169 is still buggy. What should be handled by upstream and what should be fixed is outside of this discussion and I am still available to sponsor the upload :) But btw isn't http://code.google.com/p/r8168/ the better homepage for it? Cheers -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4ebfe94d.7060...@debian.org
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
On Sun, 2011-11-13 at 15:08 +0100, Andreas Beckmann wrote: > On 2011-11-12 19:04, Patrick Matthäi wrote: > > @Andreas and Dmitry: > > You may cooperate on packaging or decide, who wants to maintain it in > > the future. > > I just revived my old ITP http://bugs.debian.org/642198 > and tried to put my things in a git repository: > Only git://git.debian.org/~anbe-guest/r8168.git seems to work so far. > > > @Ben: > > I still think this driver should be added as alternative driver to the > > archive, until r8169 will do its job for the promoted PCIIDs correctly. > > One possible solution I see would be to upload r8168 to sid (or > experimental) and add a severity=serious bug "r8168 should not enter > testing - let's fix r8169 in the kernel instead" to prevent migration. > Documentation should include detailed instructions how to properly > report bugs for the in-kernel r8169 if it does not work (but r8168 does). > > Do I see it correctly that this driver is to be classified as non-free? > The source code says "GPL2+", but 50% of the source lines is something like > mdio_write(tp, 0x1E, 0x0078); > making this look very 'blobby'. Those sequences are reconfiguring the PHY, but some of them also patch the PHY firmware. Compare with the in-tree version which uses request_firmware() for the latter. Ben. -- Ben Hutchings Never attribute to conspiracy what can adequately be explained by stupidity. signature.asc Description: This is a digitally signed message part
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
On 2011-11-12 19:04, Patrick Matthäi wrote: > @Andreas and Dmitry: > You may cooperate on packaging or decide, who wants to maintain it in > the future. I just revived my old ITP http://bugs.debian.org/642198 and tried to put my things in a git repository: Only git://git.debian.org/~anbe-guest/r8168.git seems to work so far. > @Ben: > I still think this driver should be added as alternative driver to the > archive, until r8169 will do its job for the promoted PCIIDs correctly. One possible solution I see would be to upload r8168 to sid (or experimental) and add a severity=serious bug "r8168 should not enter testing - let's fix r8169 in the kernel instead" to prevent migration. Documentation should include detailed instructions how to properly report bugs for the in-kernel r8169 if it does not work (but r8168 does). Do I see it correctly that this driver is to be classified as non-free? The source code says "GPL2+", but 50% of the source lines is something like mdio_write(tp, 0x1E, 0x0078); making this look very 'blobby'. Andreas -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4ebfcf49.5070...@abeckmann.de
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Am 10.11.2011 10:42, schrieb Dmitry Smirnov: > Package: wnpp > Severity: wishlist > Owner: Dmitry Smirnov > > * Package name: r8168-dkms > Version : 8.026.00 > Upstream Author : Realtek Semiconductor Corp. > * URL : > http://www.realtek.com.tw/downloads/downloadsView.aspx?Langid=1&PNid=13&PFid=5&Level=5&Conn=4&DownTypeID=3&GetDown=false > * License : GPL-2+ (contains binary blobs) > Programming Lang: C > Description : Realtek r8168 device driver for Linux (DKMS version) > r8168 is the Linux device driver released for RealTek RTL8168B/8111B, > RTL8168C/8111C, RTL8168CP/8111CP, RTL8168D/8111D, RTL8168DP/8111DP, and > RTL8168E/8111E Gigabit Ethernet controllers with PCI-Express interface. > . > This is to substitute built-in r8169 driver if it doesn't work well. I just searched this thread in my MUA and notified that Andreas Beckmann already filled an ITP on 20.09.2011 @ #642198 @Andreas and Dmitry: You may cooperate on packaging or decide, who wants to maintain it in the future. @Ben: I still think this driver should be added as alternative driver to the archive, until r8169 will do its job for the promoted PCIIDs correctly. -- /* Mit freundlichem Gruß / With kind regards, Patrick Matthäi GNU/Linux Debian Developer E-Mail: pmatth...@debian.org patr...@linux-dev.org */ signature.asc Description: OpenPGP digital signature
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
2011/11/10 onlyjob > Thanks, Julien. > > I see you've made r8168-source package as well - nice. > I will introduce support for module-assistant if we agree on package > inclusion. > > > I would second that as I have another problem with r8169 which doesn't > > support support resuming from suspend/hibernate. I've been using r8168 > for a > > couple of weeks without any issue. > > > > I have even built a package for my own use [0] in case this can help. > > > > I will report a bug against r8169 as advised by Ben as soon as I find the > > time to clearly describe the issue. > > > > Cheers, > > Julien > > > > [0] http://git.kirya.net/?p=debian/r8168.git;a=summary > > > -- > To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org > with a subject of "unsubscribe". Trouble? Contact > listmas...@lists.debian.org > Archive: > http://lists.debian.org/canbdodwtdz8coath0ws9b53gse3wmudef58jzlora5ss3z6...@mail.gmail.com > > BTW, I just want to say that using the r8168 driver correct the problem that I get with my network card when I reboot directly in debian from windows, without closing the computer first. It's about the WOL thing that r8169 doesn't support well or something like that.
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Thanks, Julien. I see you've made r8168-source package as well - nice. I will introduce support for module-assistant if we agree on package inclusion. > I would second that as I have another problem with r8169 which doesn't > support support resuming from suspend/hibernate. I've been using r8168 for a > couple of weeks without any issue. > > I have even built a package for my own use [0] in case this can help. > > I will report a bug against r8169 as advised by Ben as soon as I find the > time to clearly describe the issue. > > Cheers, > Julien > > [0] http://git.kirya.net/?p=debian/r8168.git;a=summary -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/canbdodwtdz8coath0ws9b53gse3wmudef58jzlora5ss3z6...@mail.gmail.com
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Hi, Le jeudi 10 nov. 2011 à 15:21:29 (+0100 CET), Patrick Matthäi a écrit : > Am 10.11.2011 15:18, schrieb Ben Hutchings: [...] > >Anyone who has problems with r8169 in 3.0 or 3.1 should report this: > > > >To: Realtek linux nic maintainers > >To: Francois Romieu > >Cc: net...@vger.kernel.org > > > >They may also wish to open a Debian bug report and cc that. > > That is right, also that it should be fixed upstream, which isn't > the case since years (according to google results). So I don't see a > problem to include the r8168-dkms package, until r8169 realy works. I would second that as I have another problem with r8169 which doesn't support support resuming from suspend/hibernate. I've been using r8168 for a couple of weeks without any issue. I have even built a package for my own use [0] in case this can help. I will report a bug against r8169 as advised by Ben as soon as I find the time to clearly describe the issue. Cheers, Julien [0] http://git.kirya.net/?p=debian/r8168.git;a=summary -- .''`. Julien Valroff ~ ~ : :' : Debian Developer & Free software contributor `. `'` http://www.kirya.net/ `- 4096R/ E1D8 5796 8214 4687 E416 948C 859F EF67 258E 26B1 -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2010191725.gc5...@kirya.net
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Am 10.11.2011 15:18, schrieb Ben Hutchings: On Thu, 2011-11-10 at 14:14 +0100, Patrick Matthäi wrote: Am 10.11.2011 13:54, schrieb onlyjob: [...] Don't get me wrong: I love alternatives, but I don't see their value when it comes to device drivers. That's probably because you can afford it due to lack of problematic hardware on your hands. For those who are not as lucky as you and me it might be more difficult. At least we already have one person who might need solution right now. ;) I just get that new NIC with my new mainboard, a Gigabyte GA-990XA-UD3-AM3. Since I just have got this hardware since 1-2 weeks, I didn't searched for other solutions, but r8168 works. btw: Linux 3.0 and 3.1 didn't worked. [...] Anyone who has problems with r8169 in 3.0 or 3.1 should report this: To: Realtek linux nic maintainers To: Francois Romieu Cc: net...@vger.kernel.org They may also wish to open a Debian bug report and cc that. That is right, also that it should be fixed upstream, which isn't the case since years (according to google results). So I don't see a problem to include the r8168-dkms package, until r8169 realy works. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4ebbdde9.7050...@debian.org
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
On Thu, 2011-11-10 at 14:14 +0100, Patrick Matthäi wrote: > Am 10.11.2011 13:54, schrieb onlyjob: [...] > >> Don't get me wrong: I love alternatives, but I don't see their value when > >> it > >> comes to device drivers. > > > > That's probably because you can afford it due to lack of problematic > > hardware on your hands. > > For those who are not as lucky as you and me it might be more difficult. > > At least we already have one person who might need solution right now. > > ;) > I just get that new NIC with my new mainboard, a Gigabyte > GA-990XA-UD3-AM3. Since I just have got this hardware since 1-2 weeks, I > didn't searched for other solutions, but r8168 works. > btw: Linux 3.0 and 3.1 didn't worked. [...] Anyone who has problems with r8169 in 3.0 or 3.1 should report this: To: Realtek linux nic maintainers To: Francois Romieu Cc: net...@vger.kernel.org They may also wish to open a Debian bug report and cc that. Ben. -- Ben Hutchings You can't have everything. Where would you put it? signature.asc Description: This is a digitally signed message part
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Am 10.11.2011 13:54, schrieb onlyjob: Dear Kai, well, then what is different and why can't that be upstreamed? And is there a known case, where the driver provided by Debian doesn't work (if so, is there a bug filed)? Perhaps you are right, given the time and the skills and the hardware and the person skilled and motivated enough it can be upstreamed. Ideally it should be. I'm afraid I don't fit for such job: I can't meet you high expectations since I might not have enough time, skills, right hardware, etc. Let's see what Patrick Matthäi is going to tell since he is apparently have a problematic hardware in his possession. r8169 "works", but I get approx every 5 minutes timeouts of one minute (where the whole NIC does not work) or I have got a package loss around 50% and it is quite slow. This is an older known problem in Debian, just search for "r8168 r8169 debian" on google. Don't get me wrong: I love alternatives, but I don't see their value when it comes to device drivers. That's probably because you can afford it due to lack of problematic hardware on your hands. For those who are not as lucky as you and me it might be more difficult. At least we already have one person who might need solution right now. ;) I just get that new NIC with my new mainboard, a Gigabyte GA-990XA-UD3-AM3. Since I just have got this hardware since 1-2 weeks, I didn't searched for other solutions, but r8168 works. btw: Linux 3.0 and 3.1 didn't worked. Apparently we have a philosophical problem here: given the fact that nether you nor me can benefit from having the alternative driver it will be over-generalization for us to say that nobody can and therefore don't give them a chance. We're not talking about mandatory driver so alternative might be good for someone. Of course it would be perfect to have one 100% working driver in kernel so all those troubles won't be necessary at all. I believe it will be like this, eventually. There was a "fix" for this problem in stable linux-2.6: 2.6.32-36: * r8169: Backport changes up to Linux 3.0.2 (Closes: #627704) - Fix support for RTL8102E and RTL8168DP - Add support for RTL8105E, RTL8168E and another variant of RTL8168DP - Add support for D-Link DGE-530T rev C1 But yeah as I told, 3.0 and 3.1 have got the same problems with my NIC. Cheers. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4ebbce43.8060...@debian.org
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
On Thu, Nov 10, 2011 at 11:54:01PM +1100, onlyjob wrote: > Perhaps you are right, given the time and the skills and the hardware > and the person skilled and motivated enough it can be upstreamed. > Ideally it should be. IIRC, upstream doesn't want r8168 because supposedly it covers only hardware that is also supported by r8169. They feel that if r8169 is buggy that it should be fixed. Obviously, that isn't helpful at the present moment if the current driver doesn't work for you. Unfortunately, I can't find a reference at the moment, but I do remember reading about it recently. -- brian m. carlson / brian with sandals: Houston, Texas, US +1 832 623 2791 | http://www.crustytoothpaste.net/~bmc | My opinion only OpenPGP: RSA v4 4096b: 88AC E9B2 9196 305B A994 7552 F1BA 225C 0223 B187 signature.asc Description: Digital signature
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Dear Kai, > well, then what is different and why can't that be upstreamed? And is there a > known case, where the driver provided by Debian doesn't work (if so, is there > a > bug filed)? Perhaps you are right, given the time and the skills and the hardware and the person skilled and motivated enough it can be upstreamed. Ideally it should be. I'm afraid I don't fit for such job: I can't meet you high expectations since I might not have enough time, skills, right hardware, etc. Let's see what Patrick Matthäi is going to tell since he is apparently have a problematic hardware in his possession. > Don't get me wrong: I love alternatives, but I don't see their value when it > comes to device drivers. That's probably because you can afford it due to lack of problematic hardware on your hands. For those who are not as lucky as you and me it might be more difficult. At least we already have one person who might need solution right now. Apparently we have a philosophical problem here: given the fact that nether you nor me can benefit from having the alternative driver it will be over-generalization for us to say that nobody can and therefore don't give them a chance. We're not talking about mandatory driver so alternative might be good for someone. Of course it would be perfect to have one 100% working driver in kernel so all those troubles won't be necessary at all. I believe it will be like this, eventually. Regards, Dmitry. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/canbdodvlyn5j7voikxcyr9bu5vrdpenjranyrjyon8q2x0s...@mail.gmail.com
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Dear Dimitry, Dmitry Smirnov schrieb am 10.11.2011 12:50: > Sorry, I don't understand... > > This is merely an alternative driver from hardware vendor. > Sometimes you might need alternatives if your hardware doesn't work well, > or if you just suspect it do not work well but you need something to > compare with. > > "Fixing stuff" is always good if you can do it. well, then what is different and why can't that be upstreamed? And is there a known case, where the driver provided by Debian doesn't work (if so, is there a bug filed)? Don't get me wrong: I love alternatives, but I don't see their value when it comes to device drivers. Kind regards, Kai Wasserbäch -- E-Mail: cu...@debian.org IRC: Curan Jabber: dri...@debianforum.de URL: http://wiki.debian.org/C%C3%B9ran GnuPG: 0xE1DE59D2 0600 96CE F3C8 E733 E5B6 1587 A309 D76C E1DE 59D2 signature.asc Description: OpenPGP digital signature
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Sorry, I don't understand... This is merely an alternative driver from hardware vendor. Sometimes you might need alternatives if your hardware doesn't work well, or if you just suspect it do not work well but you need something to compare with. "Fixing stuff" is always good if you can do it. Dmitry. On 10 November 2011 22:34, Cyril Brulebois wrote: > Hi. > > Dmitry Smirnov (10/11/2011): >> r8168 is the Linux device driver released for RealTek RTL8168B/8111B, >> RTL8168C/8111C, RTL8168CP/8111CP, RTL8168D/8111D, RTL8168DP/8111DP, and >> RTL8168E/8111E Gigabit Ethernet controllers with PCI-Express interface. >> . >> This is to substitute built-in r8169 driver if it doesn't work well. > > What happened to “upstream first”, and “fix stuff if doesn't work well”? > > Mraw, > KiBi. > -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/canbdodvxrvkzepf7gmtzr3cz9dalvlhf2gygz-uguyup8x0...@mail.gmail.com
Re: Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Hi. Dmitry Smirnov (10/11/2011): > r8168 is the Linux device driver released for RealTek RTL8168B/8111B, > RTL8168C/8111C, RTL8168CP/8111CP, RTL8168D/8111D, RTL8168DP/8111DP, and > RTL8168E/8111E Gigabit Ethernet controllers with PCI-Express interface. > . > This is to substitute built-in r8169 driver if it doesn't work well. What happened to “upstream first”, and “fix stuff if doesn't work well”? Mraw, KiBi. signature.asc Description: Digital signature
Bug#648286: ITP: r8168 -- Realtek r8168 device driver for Linux (DKMS version)
Package: wnpp Severity: wishlist Owner: Dmitry Smirnov * Package name: r8168-dkms Version : 8.026.00 Upstream Author : Realtek Semiconductor Corp. * URL : http://www.realtek.com.tw/downloads/downloadsView.aspx?Langid=1&PNid=13&PFid=5&Level=5&Conn=4&DownTypeID=3&GetDown=false * License : GPL-2+ (contains binary blobs) Programming Lang: C Description : Realtek r8168 device driver for Linux (DKMS version) r8168 is the Linux device driver released for RealTek RTL8168B/8111B, RTL8168C/8111C, RTL8168CP/8111CP, RTL8168D/8111D, RTL8168DP/8111DP, and RTL8168E/8111E Gigabit Ethernet controllers with PCI-Express interface. . This is to substitute built-in r8169 driver if it doesn't work well. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2010094245.21454.64340.reportbug@deblab
