Bug#748398: Bug has returned as CONFIG_FB_RADEON=y is set again
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Hi Ben! Here are my test results: Kernels tested: * Linux 3.16.7-ck2 and Linux 3.18.0. Hardware tested: * Apple Mac Mini G4 (Radeon 9200) * Apple Powerbook G4 (Radeon 9600 PRO Turbo) * Apple iBook G3 (Rage Mobility 128 AGP) Results before changes: * Graphics broken on Apple Mac Mini G4, both kernels, out-of-the-box, can be fixed by appending video=radeonfb:off. * Graphics broken on Apple PowerBook G4, both kernels, out-of-the-box, can be fixed by appending video=radeonfb:off. * Graphics work on Apple iBook G3, both kernels, out-of-the-box. Changes: * CONFIG_FB_RADEON=y was changed to CONFIG_FB_RADEON=m. Results: * Graphics work on Apple Mac Mini G4, both kernels, out-of-the-box, no additional kernel parameters required. * Graphics work on Apple PowerBook G4, both kernels, out-of-the-box, no additional kernel parameters required. * Graphics work on Apple iBook G3, both kernels, out-of-the-box. Remarks: * Backlight dimming control works on both laptops in this configuration. Brightness can be dimmed from 0% to 100% smoothly. No other side effects observed, graphics and display controls on both laptops and the Mac Mini work as expected. Conclusion: * Please set CONFIG_FB_RADEON=m on powerpc32. Adrian - -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -BEGIN PGP SIGNATURE- Version: GnuPG v1 iQIcBAEBCAAGBQJUtl07AAoJEHQmOzf1tfkT/CcQAK6ZQk6ITTqWqUiNhullIxXc kBqIpiGQE2UktPYbOGAtnJPb+hHypYEywei/JeCdkPSHPKYhN0SpLv3iggP4HFeL n0MKoRv3g6rQ1D5Kxysd3N8cfwKEnYvG6mjt9sP3OBmn6Z1q74+NaSfRlqWn7fqH JUsBfgRlTeqvn5SvCWG4rQMV3Kuko/McH94+9xm5M/pNs5BWNSr4CKHyZiBSIMjI BwsmHCjFpg7zKA/AES8JFPBwAJcYNSmX8ByPOcc/WaumrdB239aV9OadwNcxGVWw lBGuIcPGtxdbSS4PQdcL4aetQlKxkh9D4ka1qUS0TGvw5664gyCgjg4nPQrwPnyu veqP8igIn5eNzGauwcJ+LXfSCxaE1rNrgqoVolP/USrPosBageo7K3RSSuYPVueH HF0+BPnWmH2ZHgTOv+MV2A2ZUsuuNpxjEZPFJkSdcc99x7kKfVOQcpO3QKWhKIlg m8ExQOI23xOOgNe7HuyxH/kRHtdKouyI3n4i2/n7UcpR21q9B5yKCpDOpnjwfn7h BG8mtknkCChUCzI9ZgmMQnAHAZBGuN/a4kBM7pvWuNYP1bU73MCLZ4J+8lgw/80z mAWHtSPJ3XJL86DPL5zO7doTizet50EPcuLEMUpf77LUks7+AwUAaMAmr5ndaRCG W0P5iE7RkpcN++eax1CM =jITB -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/54b65d3b.6070...@physik.fu-berlin.de
Bug#748398: Bug has returned as CONFIG_FB_RADEON=y is set again
On Wed, 2015-01-14 at 13:12 +0100, John Paul Adrian Glaubitz wrote: Hi Ben! Here are my test results: [...] Conclusion: * Please set CONFIG_FB_RADEON=m on powerpc32. Thank you, I will. Ben. -- Ben Hutchings The world is coming to an end. Please log off. signature.asc Description: This is a digitally signed message part
Bug#748398: Bug has returned as CONFIG_FB_RADEON=y is set again
On 24 Dec 2014 00:05:02, Ben Hutchings wrote: That is still better than having no display if the backlight is not correctly controlled. I confirm that backlight is not controlled correctly. Tried kms on powerbook 5,8. With video=radeonfb:off the only thing after X loaded is a black screen. -- WBR, Vladimir Berezenko. -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/3643976.Dc8ULYr0Y0@qmhomeg5
Bug#748398: Bug has returned as CONFIG_FB_RADEON=y is set again
Hello Vladimir! On 01/04/2015 08:03 PM, Vladimir Berezenko wrote: That is still better than having no display if the backlight is not correctly controlled. I confirm that backlight is not controlled correctly. Tried kms on powerbook 5,8. With video=radeonfb:off the only thing after X loaded is a black screen. Thanks for the heads-up! Could you tell us what kernel version you used and what the version of the X.org driver was (Debian package versions)? Out of curiosity, what X.org driver are you using? Assuming you are using Debian testing with a 3.16 kernel and 7.5.0 radeon driver, you won't be able to use the radeon driver anymore but just the generic, low-performance fbdev driver. I will hopefully be able to test some Powerbooks myself this week. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/54a99384.1050...@physik.fu-berlin.de
Bug#748398: Bug has returned as CONFIG_FB_RADEON=y is set again
On 4 Jan 2015 20:24:52, John Paul Adrian Glaubitz wrote: I confirm that backlight is not controlled correctly. Tried kms on powerbook 5,8. With video=radeonfb:off the only thing after X loaded is a black screen. Thanks for the heads-up! Could you tell us what kernel version you used and what the version of the X.org driver was (Debian package versions)? I'm using debian testing versions. You are good telepathist. 8) Out of curiosity, what X.org driver are you using? Assuming you are using Debian testing with a 3.16 kernel and 7.5.0 radeon driver, you won't be able to use the radeon driver anymore but just the generic, low-performance fbdev driver. Yes. 3.16.7-ckt2-1 linux-image and 1:7.5.0-1 xserver-xorg-video-radeon. And yes, the gui is very slow and unresponsive. PS: is there any possibility to get fw for powermac g5 pci-e radeon cards? the default nvidia card is a very big pain now. -- WBR, Vladimir Berezenko. -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/1503419.mi64b5YjV3@qmhomeg5
Bug#748398: Bug has returned as CONFIG_FB_RADEON=y is set again
On Tue, 2014-12-23 at 03:24 +0100, John Paul Adrian Glaubitz wrote: On 12/23/2014 02:48 AM, Ben Hutchings wrote: This change caused a regression for some other 32-bit PowerMacs the last time we tried it (#614221). Therefore I limited it to 64-bit PowerPC configurations. Thanks for the hint! Interestingly, the linked bug report mentions that the bug showed on the Mac Mini G4 model as well [1] which is the exact model which I am testing on (G4 CPU, Radeon 9200 graphics). Also, as far as I know, the iBook G4 that you tested as well as my Mac Mini G4 share exactly the same hardware, minus the display. So I am quite confident the results should be the same. I believe backlight control was broken when using KMS, so I am less confident about that. I no longer have the iBook G4 so can't test it again. I am not going to make any further changes to this, until someone can show a solution that will not reintroduce that regression. Alright, which hardware configurations do you want me to test? As I said during DebConf, I might be able to lend some G3/G4 Macs from people at my department and run some tests on them. In any case, we need to be able to use the radeon KMS driver as Radeon X.Org driver that ships with Jessie does no longer support UMS, so anyone using the 32-bit kernel on a PowerPC will have poor graphics performance either way. That is still better than having no display if the backlight is not correctly controlled. I will try to get into touch with the people from the #614221 bug report. They should be able to test whether the regression still exists by simply adding video=radeonfb:off on a testing/sid system. Thanks. Ben. Adrian [1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=614221#39 -- Ben Hutchings Editing code like this is akin to sticking plasters on the bleeding stump of a severed limb. - me, 29 June 1999 signature.asc Description: This is a digitally signed message part
Bug#748398: Bug has returned as CONFIG_FB_RADEON=y is set again
Control: unarchive -1 Control: reopen -1 Hello Ben! I just did a dist-upgrade which also upgraded the kernel. After rebooting the machine, X was still broken as radeonfb was loaded again and blocked the radeon module from functioning properly. Since the radeonfb driver is compiled into the kernel, there is no easy way of blacklisting it except for the boot command line. It seems your changes did not actually arrive in the kernel package: root@test-adrian1:/boot grep CONFIG_FB_RADEON= config-3.16* config-3.16.0-4-powerpc:CONFIG_FB_RADEON=y config-3.16-3-powerpc:CONFIG_FB_RADEON=y root@test-adrian1:/boot Is this a mistake? It's a pretty showstopper because it forces any installation of Debian on a PowerMac with a Radeon card to use the fbdev driver by default which uses a very low colordepth and has a bad performance. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/54985637.70...@physik.fu-berlin.de
Bug#748398: Bug has returned as CONFIG_FB_RADEON=y is set again
Control: tag -1 wontfix help On Mon, 2014-12-22 at 18:34 +0100, John Paul Adrian Glaubitz wrote: Control: unarchive -1 Control: reopen -1 Hello Ben! I just did a dist-upgrade which also upgraded the kernel. After rebooting the machine, X was still broken as radeonfb was loaded again and blocked the radeon module from functioning properly. Since the radeonfb driver is compiled into the kernel, there is no easy way of blacklisting it except for the boot command line. It seems your changes did not actually arrive in the kernel package: root@test-adrian1:/boot grep CONFIG_FB_RADEON= config-3.16* config-3.16.0-4-powerpc:CONFIG_FB_RADEON=y config-3.16-3-powerpc:CONFIG_FB_RADEON=y root@test-adrian1:/boot Is this a mistake? It's a pretty showstopper because it forces any installation of Debian on a PowerMac with a Radeon card to use the fbdev driver by default which uses a very low colordepth and has a bad performance. This change caused a regression for some other 32-bit PowerMacs the last time we tried it (#614221). Therefore I limited it to 64-bit PowerPC configurations. I am not going to make any further changes to this, until someone can show a solution that will not reintroduce that regression. Ben. -- Ben Hutchings Editing code like this is akin to sticking plasters on the bleeding stump of a severed limb. - me, 29 June 1999 signature.asc Description: This is a digitally signed message part
Bug#748398: Bug has returned as CONFIG_FB_RADEON=y is set again
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 12/23/2014 02:48 AM, Ben Hutchings wrote: This change caused a regression for some other 32-bit PowerMacs the last time we tried it (#614221). Therefore I limited it to 64-bit PowerPC configurations. Thanks for the hint! Interestingly, the linked bug report mentions that the bug showed on the Mac Mini G4 model as well [1] which is the exact model which I am testing on (G4 CPU, Radeon 9200 graphics). Also, as far as I know, the iBook G4 that you tested as well as my Mac Mini G4 share exactly the same hardware, minus the display. So I am quite confident the results should be the same. I am not going to make any further changes to this, until someone can show a solution that will not reintroduce that regression. Alright, which hardware configurations do you want me to test? As I said during DebConf, I might be able to lend some G3/G4 Macs from people at my department and run some tests on them. In any case, we need to be able to use the radeon KMS driver as Radeon X.Org driver that ships with Jessie does no longer support UMS, so anyone using the 32-bit kernel on a PowerPC will have poor graphics performance either way. I will try to get into touch with the people from the #614221 bug report. They should be able to test whether the regression still exists by simply adding video=radeonfb:off on a testing/sid system. Adrian [1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=614221#39 - -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -BEGIN PGP SIGNATURE- Version: GnuPG v1 iQIcBAEBCAAGBQJUmNJNAAoJEHQmOzf1tfkT4JcP/0Ws4+rUfnFn/grdX4TrqLkt Eh18hssDnsmmqbv1CtBNHJ7XbIhCrUvxFk/8trKZwKF2cO5x7l6M8AB6SYoVxBPi xV3aLtzZMGLEu76a6a7QA7syXoFJ+V1pnWDydTlWnaL+2EyxeEVMHwe7U+Gerl+m HIPIqLn4/ICoxDTFEY+dPc6gGC3s/lSUV/YegQs5+Uh8q/hj61I0qLLMPfveTUDO DR6AzemsyG2nsr1yMUVrL6i3sXTppkXG2Mdik1DCONvfC+zoAnvAF8d8NffNioz2 aa5eQpvX9ShSrQUKhjFF5jLOBMx5iEJM1U9d1ACyIS2KjtpbwIFNWkSKIxVv2Y+Z F3Ch7jj2sWTa/M/E+3Li7/kNgN/6tUzfkg5TIDuiCdB74dfF528QHxwv7JdzysB+ P5z6RWTXbo4wESCTe2/HqT4/WvEGXx1ivPnGrKOWLAmKF0z9IKUUYzyfL0ThZowq mDxRIZwjtUPQ/lBL9XIpYqIp34CnBk0HhGf9pSkV1FKJV/MG8zD/UMCka+Eszc6d Kxf8s1JX34/forFOVtdXpR/YM9loPfb5i4CC2XHR9BdwC1Z8q2eMBMXndeSEXMjQ xgsmYJMMODNFD8unErAXp2h0fh7YSVlp+uc4wB8/VgAXCYM43ZIKw6WyC0R8bVQ+ yn3NW9kfyrGSu6C6khii =F7mk -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5498d24d.8070...@physik.fu-berlin.de
Bug#748398: Bug has returned as CONFIG_FB_RADEON=y is set again
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 12/23/2014 03:24 AM, John Paul Adrian Glaubitz wrote: On 12/23/2014 02:48 AM, Ben Hutchings wrote: This change caused a regression for some other 32-bit PowerMacs the last time we tried it (#614221). Therefore I limited it to 64-bit PowerPC configurations. Thanks for the hint! Interestingly, the linked bug report mentions that the bug showed on the Mac Mini G4 model as well [1] which is the exact model which I am testing on (G4 CPU, Radeon 9200 graphics). Oh, and to conclude: According to the linked bug report, the problem arose on the exact same hardware that I am using but no longer occurs for me. My Mac Mini G4 works absolutely fine with CONFIG_FB_RADEON=m and using the radeon module. So, my suspicion is that the bug has actually been fixed upstream. However, as I said before, I am willing to try to borrow any particular PPC Mac and test for the regression. Adrian - -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -BEGIN PGP SIGNATURE- Version: GnuPG v1 iQIcBAEBCAAGBQJUmNPCAAoJEHQmOzf1tfkTJesP/i7DmxFxIhZVtt2BImiAW9m1 /hiBiPPHF47tfqykpG9NlT1PhMVwOdcW9o00NFcZ05tXDVOxxXF+Jp4TEgfkYt5t 2SOgLoHax5lTVqBy8XJ1FR7rmjcbJEgk/cj6UXs6XQhIiWIeUfnotveKLYPDTiOQ KFEmtvKlyd7JGWcE4uUhEuUzCFxLtlTRIAZlusBkbPbXGakC854CRJCo94XTeNzJ QDe1Oj3YM94xtJlkXY2nvh8Uh6NirtfDxwATaHfZwhBlukzJ56DT1KentVu5j2o6 EJDd/b9/uFqCqFg55F1fMGWo3as3B5awRt/IJcBNCh3n9ch0mlugMotvdU4aKjIk ANUAx+e8ogG/8kg+drrGp4Wab4sBSjZhra/bLHU+o1Lzj98UKF2J1vVOTu9127uT cnPFogjJLjCgyRKy/WyU4w4IOvcrqSoC2FDoqO5vG7ZT2+9hYfyWc5kkt4up9TYp 1DPNFjpj3Ymmv1EgS4iExEXbuRQK+AYQPkpBwWhOGTyEZ1HQHBjSg7KQCO4grBRS 4GlfkLEfB6n9b14oJg5pt6kl5q7e03XxZV3byKHkjoTvp4WFnCS49G2p0aUTvqcZ u5G0wK4sbVPJytKPQN3WqscJj0IcUN15ajU5/qHQvtzVguzCqUf/qz9TKdmE68qz YRYfFWoliW7+OKHmQsVl =8UML -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5498d3c2.60...@physik.fu-berlin.de