Jump to content
Welcome to InsanelyMac Forum

Register now to gain access to all of our features. Once registered and logged in, you will be able to contribute to this site by submitting your own content or replying to existing content. You'll be able to customize your profile, receive reputation points as a reward for submitting content, while also communicating with other members via your own private inbox, plus much more! This message will be removed once you have signed in.


  • Content count

  • Joined

  • Last visited

  1. Yes. It seems deleting .IABootfiles is does the trick. Was also able to delete .IAProductInfo with no apparent problems. All looked like they'd been sitting on my hard drive for a couple of years.
  2. New Driver for Realtek RTL8111

    Thanks. (Just as a follow up fyi for anyone else with slow uploads: I looked back over thread and found at least one person suggesting using the ancient Lnx2mac RealtekRTL81xx kext on el capitan even though it's not been updated since 2014. So I'm giving that a try, it works and has full upload speed I just hope it's stable. About to stress test it by running a full backup to my home server.)
  3. New Driver for Realtek RTL8111

    Heh. That would be nice...but not really feasible right now. So you think the latency/power management issues would affect an add in pci card like the Delock just the same as the motherboard's built in GigE?
  4. New Driver for Realtek RTL8111

    I've just realised that since I upgraded to El capitan (and switched from version 1.2.3 to 2.2.1d4) I've had an issue with slow upload speeds. Transfers are actually two slow to backup to my home server over gigabit ethernet and can't even keep up to the upload rate my fibre internet can handle. My GA-ep15-ds3l mobo has an RTL8111/8168B Gigabit Ethernet (rev 02). If I boot into linux then I get the speeds I'd expect both on the LAN and WAN, so I think the issue is definitely about software and osx. I've done some testing with iperf3 on both my server and hack. I get 936 Mbits/sec (1.09GB in 10 secs) when transfering from server to hack but a measly 16.6 Mbits/sec (19.8MB in 10 secs) from hack to server – that very neatly outlines my problem! I can see no ethernet related error messages from the kernel in system.log that point to an obvious problem (but will attach those that are there below). Looking at netstat it seems that there are a lot of retransmitted packets and duplicate acks (netstat output attached below). Through umpteen reboots I've: Tried disabling various features in the kext. Tried playing with sysctl settings with only a minimal improvement. Looked again at cpu speedstep and it seems to be working, with the cpu switching states regularly. Not sure what else to look at if this is a 'power management' issue. I can see one or two others had the same or similar issues but not really any solid solutions. Any advice anyone can give on further debugging would be welcome. (@Mieze As a possible workaround does your Delock 89357 card still work on El Capitan? I had a Netgear GA311 (RTL8169 I think) for this hack when I first built it but sold it when Apple's driver stopped working on it. Happy to use a PCI GigE card if it will work.) system.log messages filtered on 'ethernet' system.log messages filtered on 'kernel'
  5. Security Update 2014-002 released

    For once nvidia have been very quick, just prompted to install web driver 331.01.01f04 and now cudaminer hash rates are back where they were plus dragging windows about when cudaminer is running isn't jerky like it was yesterday. So now everything is cool again =D
  6. Installed ok. So far: - Nvidia web driver is disabled due to not being compatible. - After starting cudaminer I had a spontaneous reboot (presume cudaminer is more compatible with nvidia drivers...)
  7. Clover General discussion

    I saw similar behaviour a couple of months ago after adding a hybrid SDHD disk to my desktop, and the install log always said everything was fine, no errors – there was problem with the Clover installer that *was* fixed. The other thing to check is that you're not booting from a different disk, USB key or whatever because of BIOS boot priority.
  8. Clover General discussion

    One idea – what sort of hard drive is it? an ssd or hybrid (sshd)? And which version of Clover are you trying to install? Until recently Clover failed to install correctly on ssd or sshd, even though the installer reports no errors, but that would leave the windows bootloader in place, hence you boot straight to windows. The issue was fixed in Clover, in early Feb, I think.
  9. New Nvidia Retail Driver Install Solution

    And today it said a CUDA update was available, downloaded and installed 5.5.43, restarted and now it's still not loading the web drivers and still saying a CUDA update is required. Weirdness. [edit] Today i prompted to install CUDA 5.5.47 which installed happily and removed the CUDA update message – but still the nvidia web drivers aren't enabled after a restart. I'm guessing something changed with regards to what the web drivers write to NVRam... It seems the web drivers are loaded, according to System Information app, just the nvidia system prefs panel that says otherwise.
  10. New Nvidia Retail Driver Install Solution

    I've been using the Apple supplied drivers for my GTX660ti since upgrading to Mavericks, but kept the nvidia pref pane installed. An ta-da! Today it just updated the nvidia web drivers to 331.01.01f01 - still CUDA is broken, presumably they'll release the latest CUDA drivers soon as well. Current issue - osx doesn't seem willing to use them and every time I log in on startup the cuda system pref pane opens up and demands an update before saying none available Perhaps these issues are related. Added nvda_drv=1 to clover kernel args, but doesn't seem to help.
  11. Clover General discussion

    Or even more easily by running the Clover installer I was always working from a separate drive, and unmounting the efi partition (you can't unmount a drive, just its partitions) - at that point I was probably just issuing the wrong commands though...my bad debugging technique as always, changing too many variables Using a USB stick was another suggested alternative but I was a bit remiss about keeping my sticks Clover version up to date and the old version was causing issues with restart/shutdown, requiring hard reset and generally making things worse. I've updated it now I'm up and running again. Yes, and as I understand it the reasoning behind it is: the partition geometry is in the first 512 bytes and if you just write boot1 straight to the sector you erase the geometry and end up with a corrupted partition map. (I did this a couple of times.) So you copy out the original, create a new one with boot1 added then append the geometry info to it, before finally copying the lot back into place. True - but that still means any new versions of boot0 and 1 won't get applied unless its done manually, whilst booted from another system etc. Which brings me to a final question: does the Clover installer need fixed to sort this? Can it be? Increasingly these 4k discs (if that's the root of the problem and not the 'hybrid') will be ALL hard discs. I know the installer mounts the efi partition - is that perhaps the problem? The MBR and PBR need written before its mounted? I have no idea how the installer currently works =D
  12. Clover General discussion

    Success with disk utility format and partition, followed by linux manual install of boot0af and boot1f32. No need to specify bs=4096, presumably because the drive firmware is emulating 512b sectors anyway. To summarize for anyone else that has a boot0af: error after running clover installer on a Seagate hybrid SSHD serial ST1000DX001 aiming for legacy bios boot. First partitioned the disk with disk utility so there was at least the osx efi partition and an hfs+ partition to install OSX into. Second I ran the clover installer and pointed it towards that new hfs+ partition on the hybrid. Third, with clover having previously installed on another non-hybrid disk linux sees as /dev/sdb3, I was able to do the following in parted magic (1): Linux is seeing the hybrid disk as /dev/sda in my case, and the efi partition on that as /dev/sda1 dd if=/dev/sda count=1 bs=512 of=origMBR cp origMBR newMBR dd if=/media/sdb3/usr/standalone/i386/boot0af of=newMBR bs=1 count=440 conv=notrunc dd if=newMBR of=/dev/sda count=1 bs=512 dd if=/dev/sda1 count=1 bs=512 of=origbs cp /media/sdb3/usr/standalone/i386/boot1f32 newbs dd if=origbs of=newbs skip=3 seek=3 bs=1 count=87 conv=notrunc dd if=newbs of=/dev/sda1 count=1 bs=512 To install boot0af and boot1f32 respectively. (1) you could also extract the boot0 and boot1 files you need and put them on a usb, anything you can read from parted magic once you're booted into that. BUT I'm not sure its worth the effort for me. I presume that now every time I want to upgrade Clover with the installer it will screw up the PBR and I'll have to manually upgrade with linux? Might be more sensible for me to just use the non-hybrid disk for boot with clover and shift the grub2 boot partition currently on there onto the hybrid instead...or at least less hassle. EDIT Spoke too soon! The hybrid drive is now not even responding to bios, completely invisible. Sigh. Giving up trying to boot osx from it, thats for sure. Edit2 And came back after a full power off and on. It seems the hybrid drive will sometimes fail to restart properly if you need to do a hardware reset. No harm, just not terribly reliable as a technology I think.
  13. Clover General discussion

    Positive progress, just managed two successful boots. So the actual commands on that thread are out of date, and apply to chameleon in a normal HFS+ partition not EFI, using them on EFI leeds to corrupt EFI partition, no surprise really. I just erased the hybrid disk again, and partitioned with gparted (parted magic version) - which leaves the EFI partition as fat32 and isn't recognised as a proper OSX EFI partition but still works. Then did a full linux manual install of boot0af and boot1f32 as per clover wiki installation section. Finally, restored my working EFI partition to the new one with disk utility (to copy over the other clover files, fakesmc and my config.plist). And it works at last. About to try one last thing, and repartition the disk with Disk Utility before doing a another linux manual install of boot0/1, should work I think.
  14. Clover General discussion

    Yes, it's a Seagate ST1000DX001 - 4096 Bytes per sector (4K physical emulated at 512-byte sectors) as the manual says. I noted posts suggesting that boot1 needs manually copied into place specifying 4k sector size so I tried that already (made no difference): That last line of 512 bytes transferred in seemed odd since we started copying out 4096 so I repeated it specifying 4096: dd if=/tmp/newbs of=/dev/rdisk1s1 count=1 bs=4096 But the result still said '512 bytes transferred'! Maybe I've done something wrong there, but 'man dd' doesn't suggest anything to me. Here's the clover install log fwiw:
  15. Clover General discussion

    Yes. The problem is that the clover install log shows boot1f32 being installed to the hybrid drives efi partition apparently successfully, no errors. I also tried a manual install of boot1f32 to the hybrid drives efi partition, again, it appears to work perfectly, no errors. But no matter what I do I get boot0af: error when I try to boot from the hybrid drive. The same clover installer works perfectly on a second non-hybrid hard drive - that at least gives me a workaround, boot using the second drive and just select OSX installed on the hybrid from clover UI, but I just don't see why I can't boot from the hybrid. Given the fight I've been having I think the next thing to do might be to erase the hybrid and repartition again for a fresh start just in case the efi partition or PBR is corrupted some how. I've been trying to avoid that because it means another ~6hr restore from time machine...