Announcement

Collapse
No announcement yet.

Intel GPU Driver Tries To Rip Out FBDEV Support

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #31
    Originally posted by robclark View Post
    But VT switch is essentially in userspace today. Or rather it depends on userspace for the hand-off from X. So it is already as unreliable as it's going to get.

    *cough* thank you *cough*

    Comment


    • #32
      Originally posted by Ericg View Post
      Interesting, any particular reason it doesnt work well with Gentoo? you've piqued my curiosity haha
      http://dev.gentoo.org/~aidecoe/doc/en/plymouth.xml

      Try it out. Last time I tried I experienced an insane amount of screen flickering, and the plymouth process would hang about three quarters through.

      Comment


      • #33
        Originally posted by duby229 View Post
        http://dev.gentoo.org/~aidecoe/doc/en/plymouth.xml

        Try it out. Last time I tried I experienced an insane amount of screen flickering, and the plymouth process would hang about three quarters through.
        I would but I don't keep a Gentoo installation around anymore, compiling the full kernel on my laptop alone takes an hour so I'm not jumping at compiling -everything-, every update.

        Flickering was probably a kms issue... initramfs compiled with your driver module? Plymouth hanging could have just been a bug or a symptom of the kms flickering. Personally I'd just give it a fresh shot if you haven't recently and see how it works out-- from both stable and git see if either works.

        Comment


        • #34
          Originally posted by Ericg View Post
          I would but I don't keep a Gentoo installation around anymore, compiling the full kernel on my laptop alone takes an hour so I'm not jumping at compiling -everything-, every update.

          Flickering was probably a kms issue... initramfs compiled with your driver module? Plymouth hanging could have just been a bug or a symptom of the kms flickering. Personally I'd just give it a fresh shot if you haven't recently and see how it works out-- from both stable and git see if either works.
          I don't know about Plymouth on Gentoo, but if your laptop takes an hour to build the kernel either you're doing it wrong or you have some truly pathetic hardware. On an AMD Neo @800 MHz (modprobe cpufreq_powersave, for thermal reasons), with a single thread (Debian Squeeze, GCC 4.4) I'm getting under 30 minutes. That is a custom config, but with make localmodconfig there's no reason not to make one.

          Comment


          • #35
            Originally posted by Ibidem View Post
            I don't know about Plymouth on Gentoo, but if your laptop takes an hour to build the kernel either you're doing it wrong or you have some truly pathetic hardware. On an AMD Neo @800 MHz (modprobe cpufreq_powersave, for thermal reasons), with a single thread (Debian Squeeze, GCC 4.4) I'm getting under 30 minutes. That is a custom config, but with make localmodconfig there's no reason not to make one.
            Sandy Bridge i5, dual core plus hyperthreading, clocked at ~1.6ghz i think (its a ULV model because its an ultrabook) 4Gbs of RAM, full upstream 3.9.5 kernel with the default fedora config, make -j4

            Takes me about an hour. Open to suggestions (on Arch it would take me 90mins to 2hrs, same laptop)

            Comment


            • #36
              Originally posted by Ericg View Post
              Sandy Bridge i5, dual core plus hyperthreading, clocked at ~1.6ghz i think (its a ULV model because its an ultrabook) 4Gbs of RAM, full upstream 3.9.5 kernel with the default fedora config, make -j4

              Takes me about an hour. Open to suggestions (on Arch it would take me 90mins to 2hrs, same laptop)
              That's insane. You don't have to compile everything. I just compile what I use and it takes literally like 5 minutes.

              EDIT: actually 2m27s to be exact... If you consider that you have a dual core and mine is quad double that to roughly 5 minutes. Additionally my chip is 3.2ghz so double that again to 10minutes. I'm guessing that if you stripped the kernel config down to only what you use you could compile a kernel on your system in about 10 minutes.
              Last edited by duby229; 06-19-2013, 11:08 PM.

              Comment


              • #37
                Originally posted by duby229 View Post
                That's insane. You don't have to compile everything. I just compile what I use and it takes literally like 5 minutes.

                full kernel recompile was the easiest thing to automate, personally all I really NEED is to recompile the Intel driver so that I can apply a patch after every kernel update but meh, I usually just let it run overnight
                Last edited by Ericg; 06-19-2013, 11:09 PM.

                Comment


                • #38
                  Just for my own curiosity I did a full compile of the kernel... make allyesconfig && time make -j4 bzImage modules.....real 25m24.944s

                  Comment


                  • #39
                    Originally posted by duby229 View Post
                    Just for my own curiosity I did a full compile of the kernel... make allyesconfig && time make -j4 bzImage modules.....real 25m24.944s
                    I'll wipe ccache's...cache and rerun a compile tomorrow and let ya know how long it took.

                    Now that i'm thinking about it, i wonder if running it yakuake affects its priority, hmmm..

                    Comment


                    • #40
                      ccache? I don't know where the user Gentoo wiki went, but there it was quite clearly shown that generally it's not a good idea to use it. For kernel compilation it should make even less sense, since the cache is already managed by make.

                      Comment


                      • #41
                        Originally posted by Ericg View Post
                        Sandy Bridge i5, dual core plus hyperthreading, clocked at ~1.6ghz i think (its a ULV model because its an ultrabook) 4Gbs of RAM, full upstream 3.9.5 kernel with the default fedora config, make -j4

                        Takes me about an hour. Open to suggestions (on Arch it would take me 90mins to 2hrs, same laptop)
                        Are you building the kernel using SRPM or directly from a kernel tree?
                        If you're building an SRPM, make sure you .rpmmacros is configured correctly (%_smp_mflags).

                        - Gilboa
                        DEV: Intel S2600C0, 2xE52658V2, 32GB, 4x2TB + 2x3TB, GTX780, F21/x86_64, Dell U2711.
                        SRV: Intel S5520SC, 2xX5680, 36GB, 4x2TB, GTX550, F21/x86_64, Dell U2412..
                        BACK: Tyan Tempest i5400XT, 2xE5335, 8GB, 3x1.5TB, 9800GTX, F21/x86-64.
                        LAP: ASUS N56VJ, i7-3630QM, 16GB, 1TB, 635M, F21/x86_64.

                        Comment


                        • #42
                          Originally posted by gilboa View Post
                          Are you building the kernel using SRPM or directly from a kernel tree?
                          If you're building an SRPM, make sure you .rpmmacros is configured correctly (%_smp_mflags).

                          - Gilboa
                          Direct kernel tree. Make clean, wiped ~/.ccache, removed ccache and redid the compile. Times are below.

                          Code:
                          [eric@eric-laptop linux-3.9.5]$ time make oldconfig > ~/makeoldconfig2.log
                          
                          real    0m3.528s
                          user    0m2.715s
                          sys     0m0.465s
                          
                          [eric@eric-laptop linux-3.9.5]$ time make bzImage > ~/makebzImage2.log
                          /home/eric/Source/linux-3.9.5/usr/include/linux/kexec.h:49: userspace cannot reference function or variable defined in the kernel
                          /home/eric/Source/linux-3.9.5/usr/include/linux/soundcard.h:1054: userspace cannot reference function or variable defined in the kernel
                          Succeed: decoded and checked 1692280 instructions
                          Setup is 16940 bytes (padded to 17408 bytes).
                          System is 4932 kB
                          CRC 61e8544a
                          
                          real    14m38.881s
                          user    13m10.536s
                          sys     1m36.039s
                          
                          [eric@eric-laptop linux-3.9.5]$ time make modules > ~/makemodules2.log
                          drivers/isdn/hardware/mISDN/hfcpci.c: In function ‘hfcpci_softirq’:
                          drivers/isdn/hardware/mISDN/hfcpci.c:2298:2: warning: ignoring return value of ‘driver_for_each_device’, declared with attribute warn_unused_result [-Wunused-result]
                          drivers/md/dm-raid.c: In function ‘raid_ctr’:
                          drivers/md/dm-raid.c:427:54: warning: ‘rebuilds_per_group’ may be used uninitialized in this function [-Wmaybe-uninitialized]
                          drivers/md/dm-raid.c:383:11: note: ‘rebuilds_per_group’ was declared here
                          drivers/media/usb/dvb-usb/m920x.c: In function ‘m920x_probe’:
                          drivers/media/usb/dvb-usb/m920x.c:91:6: warning: ‘ret’ may be used uninitialized in this function [-Wuninitialized]
                          drivers/media/usb/dvb-usb/m920x.c:70:6: note: ‘ret’ was declared here
                          drivers/scsi/advansys.c:71:2: warning: #warning this driver is still not properly converted to the DMA API [-Wcpp]
                          drivers/scsi/aic94xx/aic94xx_sds.c: In function ‘asd_read_flash’:
                          drivers/scsi/aic94xx/aic94xx_sds.c:597:21: warning: ‘offs’ may be used uninitialized in this function [-Wmaybe-uninitialized]
                          drivers/scsi/aic94xx/aic94xx_sds.c:985:6: note: ‘offs’ was declared here
                          drivers/scsi/mvsas/mv_sas.c: In function ‘mvs_update_phyinfo’:
                          drivers/scsi/mvsas/mv_sas.c:1162:34: warning: comparison between ‘enum sas_device_type’ and ‘enum sas_dev_type’ [-Wenum-compare]
                          drivers/scsi/mvsas/mv_sas.c:1165:39: warning: comparison between ‘enum sas_device_type’ and ‘enum sas_dev_type’ [-Wenum-compare]
                          
                          real    64m3.603s
                          user    57m23.668s
                          sys     6m59.454s
                          
                          [eric@eric-laptop linux-3.9.5]$
                          That was standard 'make' with default number of jobs (1?).

                          Make -j4 is next, i'll edit or repost with those numbers.

                          Comment


                          • #43
                            make clean, reran oldconfig,bzimage and modules...

                            Code:
                            [eric@eric-laptop linux-3.9.5]$ time make -j4 oldconfig > ~/makeoldconfig3.log
                            
                            real    0m3.520s
                            user    0m3.204s
                            sys     0m0.392s
                            
                            [eric@eric-laptop linux-3.9.5]$ time make -j4 bzImage > ~/makebzImage3.log
                            /home/eric/Source/linux-3.9.5/usr/include/linux/kexec.h:49: userspace cannot reference function or variable defined in the kernel
                            /home/eric/Source/linux-3.9.5/usr/include/linux/soundcard.h:1054: userspace cannot reference function or variable defined in the kernel
                            Succeed: decoded and checked 1692280 instructions
                            Setup is 16940 bytes (padded to 17408 bytes).
                            System is 4932 kB
                            CRC 2e1234a2
                            
                            real    7m16.351s
                            user    22m4.562s
                            sys     1m58.334s
                            
                            [eric@eric-laptop linux-3.9.5]$ time make -j4 modules > ~/makemodules3.log
                            drivers/isdn/hardware/mISDN/hfcpci.c: In function ‘hfcpci_softirq’:
                            drivers/isdn/hardware/mISDN/hfcpci.c:2298:2: warning: ignoring return value of ‘driver_for_each_device’, declared with attribute warn_unused_result [-Wunused-result]
                            drivers/md/dm-raid.c: In function ‘raid_ctr’:
                            drivers/md/dm-raid.c:427:54: warning: ‘rebuilds_per_group’ may be used uninitialized in this function [-Wmaybe-uninitialized]
                            drivers/md/dm-raid.c:383:11: note: ‘rebuilds_per_group’ was declared here
                            drivers/media/usb/dvb-usb/m920x.c: In function ‘m920x_probe’:
                            drivers/media/usb/dvb-usb/m920x.c:91:6: warning: ‘ret’ may be used uninitialized in this function [-Wuninitialized]
                            drivers/media/usb/dvb-usb/m920x.c:70:6: note: ‘ret’ was declared here
                            drivers/scsi/advansys.c:71:2: warning: #warning this driver is still not properly converted to the DMA API [-Wcpp]
                            drivers/scsi/aic94xx/aic94xx_sds.c: In function ‘asd_read_flash’:
                            drivers/scsi/aic94xx/aic94xx_sds.c:597:21: warning: ‘offs’ may be used uninitialized in this function [-Wmaybe-uninitialized]
                            drivers/scsi/aic94xx/aic94xx_sds.c:985:6: note: ‘offs’ was declared here
                            drivers/scsi/mvsas/mv_sas.c: In function ‘mvs_update_phyinfo’:
                            drivers/scsi/mvsas/mv_sas.c:1162:34: warning: comparison between ‘enum sas_device_type’ and ‘enum sas_dev_type’ [-Wenum-compare]
                            drivers/scsi/mvsas/mv_sas.c:1165:39: warning: comparison between ‘enum sas_device_type’ and ‘enum sas_dev_type’ [-Wenum-compare]
                            
                            real    29m54.490s
                            user    101m40.549s
                            sys     8m56.427s
                            [eric@eric-laptop linux-3.9.5]$
                            Odd... I seem to remember this taking longer the first time I did it just a month or so ago...

                            Either i timed it with 1 job, there was something else hogging the CPU, (This was run with me still using firefox occasionally, it had the CPU to itself most of the time when I ran it but not all the time), or having yakuake not visible decreases the priority of whatever is running and therefore slows down the compile.

                            Comment


                            • #44
                              Thats not bad for the hardware you have. I think you can still improve it a bit though.

                              Comment


                              • #45
                                Originally posted by duby229 View Post
                                Thats not bad for the hardware you have. I think you can still improve it a bit though.
                                Ideas for how? I've always just put up with it but I'd be open to improvements.

                                Comment

                                Working...
                                X