Arch Linux

ngoonee

Forum Fellow

From: Between Thailand and Singapore

Registered: 2009-03-17

Posts: 7,358

Re: Terrible performance regression with Nvidia 390.25 driver

I hope Arch team doesn't just package Nvidia's stuff without doing lots of checks. You just can't trust them.

When in doubt, follow upstream. The real solution here is not to rely on nvidia, rather than hoping each distro rely on some older version than the currently 'stable' release based on some criteria.

---

Allan-Volunteer on the (topic being discussed) mailn lists. You never get the people who matters attention on the forums.
jasonwryan-Installing Arch is a measure of your literacy. Maintaining Arch is a measure of your diligence. Contributing to Arch is a measure of your competence.
Griemak-Bleeding edge, not bleeding flat. Edge denotes falls will occur from time to time. Bring your own parachute.

kokoko3k

Member

Registered: 2008-11-14

Posts: 2,421

Re: Terrible performance regression with Nvidia 390.25 driver

Ugh, even this Chromium perf fix was due to Chromium team. Nvidia made a lot of changes with their pipeline and notified no one and just pushed it onto everyone.

As far as i understood, pipeline changes has nothing to do with denied file access permissions by the chrome(ium) sandbox.

---

Help me to improve ssh-rdp !
Retroarch User? Try my koko-aio shader !

Tom B

Member

Registered: 2014-01-15

Posts: 187

Website

Re: Terrible performance regression with Nvidia 390.25 driver

The previous nvidia driver are almost never compatibility with a next major kernel release due to changes in it.

Off topic but why is this? It feels like very poor design from the Linux kernel devs breaking backwards compatibility all the time. It gives driver developers constant work. I'd get annoyed if I had to re-write parts of a working driver every time a new kernel version was released.

V1del

Forum Moderator

Registered: 2012-10-16

Posts: 23,231

Re: Terrible performance regression with Nvidia 390.25 driver

Right now there are a few more invasive changes than usual happening wrt meltdown/spectre and co, usually relevant adjustments can be done without a change from nvidia's side (the actual kernel module, or at least the parts interfacing with the kernel is open source, however there's only a small part which then has to hook back into the proprietary parts)

And in general the kernel devs can do it because all drivers they care about are in tree and will be adjusted simultaneously as necessary.

However nvidia is usually quite quick on the pick up and it doesn't take more than a few weeks for compatibility to arrive. And Arch in general doesn't push the first kernel releases to stable anyway there are often enough small issues in the .0 releases regardless of the nvidia driver being used or not.

loqs

Member

Registered: 2014-03-06

Posts: 18,053

Re: Terrible performance regression with Nvidia 390.25 driver

@Tom B stable-api-nonsense.rst

Batou

Member

Registered: 2017-01-03

Posts: 259

Re: Terrible performance regression with Nvidia 390.25 driver

Ugh, even this Chromium perf fix was due to Chromium team. Nvidia made a lot of changes with their pipeline and notified no one and just pushed it onto everyone.

As far as i understood, pipeline changes has nothing to do with denied file access permissions by the chrome(ium) sandbox.

This whole mess is due to how memory is allocated for the graphics pipeline by the new driver. They (Nvidia) don't say it anywhere in changelogs, but it has to be due to Spectre mitigations. They kept quiet about it because of bad PR (and lower stock prices) associated with Spectre-affected products and companies. You can find some more info in the Chromium bug tracker link that I posted on the previous page.

Last edited by Batou (2018-04-06 03:24:30)

---

Please vote for all the AUR packages you're using. You can mass-vote for all of them by doing: "pacman -Qqm | xargs aurvote -v" (make sure to run "aurvote --configure"  first)

loqs

Member

Registered: 2014-03-06

Posts: 18,053

Re: Terrible performance regression with Nvidia 390.25 driver

When in doubt, follow upstream. The real solution here is not to rely on nvidia, rather than hoping each distro rely on some older version than the currently 'stable' release based on some criteria.

Reverting the upstream change does not seem to match your suggested solution.
https://git.archlinux.org/svntogit/pack … 03828c866a

ngoonee

Forum Fellow

From: Between Thailand and Singapore

Registered: 2009-03-17

Posts: 7,358

Re: Terrible performance regression with Nvidia 390.25 driver

When in doubt, follow upstream. The real solution here is not to rely on nvidia, rather than hoping each distro rely on some older version than the currently 'stable' release based on some criteria.

Reverting the upstream change does not seem to match your suggested solution.
https://git.archlinux.org/svntogit/pack … 03828c866a

I think 'follow upstream' is the reason a specific revert is done rather than a downgrade version wise (or a freeze, which is what seems to be requested in the comment I replied to).

---

Allan-Volunteer on the (topic being discussed) mailn lists. You never get the people who matters attention on the forums.
jasonwryan-Installing Arch is a measure of your literacy. Maintaining Arch is a measure of your diligence. Contributing to Arch is a measure of your competence.
Griemak-Bleeding edge, not bleeding flat. Edge denotes falls will occur from time to time. Bring your own parachute.

loqs

Member

Registered: 2014-03-06

Posts: 18,053

Re: Terrible performance regression with Nvidia 390.25 driver

I think 'follow upstream' is the reason a specific revert is done rather than a downgrade version wise (or a freeze, which is what seems to be requested in the comment I replied to).

Ah I took follow upstream to mean more upstream has made a decision lets follow it or ask upstream to change that decision rather than just locally reverting it and not working with upstream.

Kabir

Member

From: India

Registered: 2016-12-06

Posts: 59

Re: Terrible performance regression with Nvidia 390.25 driver

Can those affected test if chromium 65.0.3325.181-5 which has a patch for the sandbox interaction with the nvidia driver resolves that issue?

Chromium isnt crashing anymore upon resume from suspend or hibernation, which is awesome. Thanks so much!

Tom B

Member

Registered: 2014-01-15

Posts: 187

Website

Re: Terrible performance regression with Nvidia 390.25 driver

The last few 65.0 releases have fixed the issue, but with 66.0.3359.117 I am getting stuttering while scrolling down some pages and slow to respond tabs once again. It's definitely not as bad as it was when this thread was opened but there's a noticeable regression between  65.0.3325.181-5 and 66.0.3359.117.

blispx

Member

Registered: 2017-11-29

Posts: 53

Re: Terrible performance regression with Nvidia 390.25 driver

Strange, this problem does not occur with me

There is another annoying problem:

[    7.468397] ------------[ cut here ]------------
[    7.468399] Bad or missing usercopy whitelist? Kernel memory exposure attempt detected from SLUB object 'nvidia_stack_cache' (offset 11440, size 3)!
[    7.468407] WARNING: CPU: 2 PID: 624 at mm/usercopy.c:81 usercopy_warn+0x7e/0xa0
[    7.468407] Modules linked in: snd_hda_codec_hdmi(+) snd_hda_codec_realtek nct6775 hwmon_vid snd_hda_codec_generic arc4 ath9k ath9k_common ath9k_hw ath nls_iso8859_1 nls_cp437 vfat fat ofpart cmdlinepart intel_spi_platform eeepc_wmi intel_spi asus_wmi spi_nor sparse_keymap wmi_bmof mac80211 mtd snd_hda_intel iTCO_wdt iTCO_vendor_support snd_hda_codec intel_rapl ath3k x86_pkg_temp_thermal intel_powerclamp btusb coretemp kvm_intel btrtl btbcm btintel snd_hda_core kvm bluetooth irqbypass snd_hwdep intel_cstate e1000e intel_uncore snd_pcm intel_rapl_perf pcspkr i2c_i801 cfg80211 input_leds snd_timer mei_me ptp snd ecdh_generic led_class rfkill joydev mousedev lpc_ich pps_core mei soundcore wmi shpchp evdev rtc_cmos mac_hid sg crypto_user ip_tables x_tables ext4 crc16 mbcache jbd2 fscrypto algif_skcipher
[    7.468430]  af_alg dm_crypt dm_mod hid_generic usbhid hid sd_mod crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel pcbc ahci libahci xhci_pci ehci_pci xhci_hcd aesni_intel ehci_hcd libata aes_x86_64 crypto_simd glue_helper scsi_mod cryptd usbcore usb_common nvidia_drm(PO) drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops drm agpgart nvidia_uvm(PO) nvidia_modeset(PO) nvidia(PO) ipmi_devintf ipmi_msghandler
[    7.468443] CPU: 2 PID: 624 Comm: gnome-shell Tainted: P           O     4.16.2-2-ARCH #1
[    7.468443] Hardware name: ASUS All Series/Z87I-PRO, BIOS 1005 11/26/2014
[    7.468445] RIP: 0010:usercopy_warn+0x7e/0xa0
[    7.468445] RSP: 0018:ffffa6e1c29e7b58 EFLAGS: 00010286
[    7.468446] RAX: 0000000000000000 RBX: ffff9b1636655cb0 RCX: 0000000000000001
[    7.468447] RDX: 0000000080000001 RSI: ffffffff84e67e3c RDI: 00000000ffffffff
[    7.468448] RBP: 0000000000000003 R08: 0000000000000098 R09: 000000000000033b
[    7.468448] R10: ffffffff84ea4121 R11: 0000000000000001 R12: 0000000000000001
[    7.468449] R13: ffff9b1636655cb3 R14: 0000000000000000 R15: ffff9b1636655cf8
[    7.468450] FS:  00007ff53cd92a00(0000) GS:ffff9b165ed00000(0000) knlGS:0000000000000000
[    7.468450] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    7.468451] CR2: 0000559fb185ce38 CR3: 00000003f64ce006 CR4: 00000000001606e0
[    7.468452] Call Trace:
[    7.468455]  __check_object_size+0x130/0x1a0
[    7.468557]  os_memcpy_to_user+0x21/0x40 [nvidia]
[    7.468670]  _nv009377rm+0xbf/0xe0 [nvidia]
[    7.468760]  ? _nv028067rm+0x79/0x90 [nvidia]
[    7.468850]  ? _nv028067rm+0x55/0x90 [nvidia]
[    7.468932]  ? _nv013694rm+0xee/0x100 [nvidia]
[    7.469015]  ? _nv015342rm+0x154/0x270 [nvidia]
[    7.469126]  ? _nv008310rm+0x134/0x1a0 [nvidia]
[    7.469237]  ? _nv008289rm+0x29c/0x2b0 [nvidia]
[    7.469347]  ? _nv001072rm+0xe/0x20 [nvidia]
[    7.469459]  ? _nv007316rm+0xd8/0x100 [nvidia]
[    7.469568]  ? _nv001171rm+0x627/0x830 [nvidia]
[    7.469677]  ? rm_ioctl+0x73/0x100 [nvidia]
[    7.469732]  ? nvidia_ioctl+0x573/0x720 [nvidia]
[    7.469734]  ? kmem_cache_free+0x1bd/0x1f0
[    7.469788]  ? nvidia_frontend_unlocked_ioctl+0x3e/0x50 [nvidia]
[    7.469790]  ? do_vfs_ioctl+0xa4/0x630
[    7.469792]  ? __fput+0x131/0x1e0
[    7.469793]  ? SyS_ioctl+0x74/0x80
[    7.469795]  ? do_syscall_64+0x74/0x190
[    7.469798]  ? entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[    7.469799] Code: 48 c7 c0 e9 20 e7 84 48 0f 44 c2 41 50 51 41 51 48 89 f9 49 89 f1 4d 89 d8 4c 89 d2 48 89 c6 48 c7 c7 40 21 e7 84 e8 02 a7 e3 ff <0f> 0b 48 83 c4 18 c3 48 c7 c6 3c 7e e6 84 49 89 f1 49 89 f3 eb 
[    7.469818] ---[ end trace ffba71c95aa2554a ]---

I am waiting for a new kernel/or nvidia package

Last edited by blispx (2018-04-19 22:22:53)

loqs

Member

Registered: 2014-03-06

Posts: 18,053

Re: Terrible performance regression with Nvidia 390.25 driver

@blispx that is expected with 4.16 https://bugs.archlinux.org/task/58074

MaddyBoo

Member

From: Portland, OR

Registered: 2017-05-02

Posts: 10

Website

Re: Terrible performance regression with Nvidia 390.25 driver

Can anyone confirm this is mitigated in kernel 4.16?

I've stuck with nvidia 387 for the past 2 months but apparently it is incompatible with kernel 4.16, I don't want to upgrade and break my system.

Batou

Member

Registered: 2017-01-03

Posts: 259

Re: Terrible performance regression with Nvidia 390.25 driver

Can anyone confirm this is mitigated in kernel 4.16?

I've stuck with nvidia 387 for the past 2 months but apparently it is incompatible with kernel 4.16, I don't want to upgrade and break my system.

Everything works fine now. Biggest issue was Chromium/Chrome. Now that works well too. If you get a black screen after upgrade, clean up xorg.conf and remove anything related to pipeline. You can even remove xorg.conf completely since Xorg doesn't need it for default options anymore.

---

Please vote for all the AUR packages you're using. You can mass-vote for all of them by doing: "pacman -Qqm | xargs aurvote -v" (make sure to run "aurvote --configure"  first)

phunni

Member

From: Bristol, UK

Registered: 2003-08-13

Posts: 787

Re: Terrible performance regression with Nvidia 390.25 driver

Updating everything that's been stuck on an older version now... Somewhat nervous!

Edit: Some very quick and basic testing later - everything looks fine! Huzzah!

Last edited by phunni (2018-04-20 13:16:34)

cirrus9

Member

Registered: 2016-04-15

Posts: 51

Re: Terrible performance regression with Nvidia 390.25 driver

Everything is working well on my system now with linux 4.16.3, and nvidia 390.48-9. Let's hope it stays that way for a while...

blispx

Member

Registered: 2017-11-29

Posts: 53

Re: Terrible performance regression with Nvidia 390.25 driver

@loqs thanks

I will build 396 and see if they have improved with a patch

Tom B

Member

Registered: 2014-01-15

Posts: 187

Website

Re: Terrible performance regression with Nvidia 390.25 driver

The last few 65.0 releases have fixed the issue, but with 66.0.3359.117 I am getting stuttering while scrolling down some pages and slow to respond tabs once again. It's definitely not as bad as it was when this thread was opened but there's a noticeable regression between  65.0.3325.181-5 and 66.0.3359.117.

This is really annoying, while typing into a  rich text area e.g. for emails in gmail or various websites it will freeze for a few seconds. The more complex the page, the worse the results. Even running --no-sandbox makes no difference. This was fine until 66.0 and 4.16.

Is there anything I can do to debug this?

Batou

Member

Registered: 2017-01-03

Posts: 259

Re: Terrible performance regression with Nvidia 390.25 driver

Strange, this problem does not occur with me

There is another annoying problem:

[    7.468397] ------------[ cut here ]------------
[    7.468399] Bad or missing usercopy whitelist? Kernel memory exposure attempt detected from SLUB object 'nvidia_stack_cache' (offset 11440, size 3)!
[    7.468407] WARNING: CPU: 2 PID: 624 at mm/usercopy.c:81 usercopy_warn+0x7e/0xa0
[    7.468407] Modules linked in: snd_hda_codec_hdmi(+) snd_hda_codec_realtek nct6775 hwmon_vid snd_hda_codec_generic arc4 ath9k ath9k_common ath9k_hw ath nls_iso8859_1 nls_cp437 vfat fat ofpart cmdlinepart intel_spi_platform eeepc_wmi intel_spi asus_wmi spi_nor sparse_keymap wmi_bmof mac80211 mtd snd_hda_intel iTCO_wdt iTCO_vendor_support snd_hda_codec intel_rapl ath3k x86_pkg_temp_thermal intel_powerclamp btusb coretemp kvm_intel btrtl btbcm btintel snd_hda_core kvm bluetooth irqbypass snd_hwdep intel_cstate e1000e intel_uncore snd_pcm intel_rapl_perf pcspkr i2c_i801 cfg80211 input_leds snd_timer mei_me ptp snd ecdh_generic led_class rfkill joydev mousedev lpc_ich pps_core mei soundcore wmi shpchp evdev rtc_cmos mac_hid sg crypto_user ip_tables x_tables ext4 crc16 mbcache jbd2 fscrypto algif_skcipher
[    7.468430]  af_alg dm_crypt dm_mod hid_generic usbhid hid sd_mod crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel pcbc ahci libahci xhci_pci ehci_pci xhci_hcd aesni_intel ehci_hcd libata aes_x86_64 crypto_simd glue_helper scsi_mod cryptd usbcore usb_common nvidia_drm(PO) drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops drm agpgart nvidia_uvm(PO) nvidia_modeset(PO) nvidia(PO) ipmi_devintf ipmi_msghandler
[    7.468443] CPU: 2 PID: 624 Comm: gnome-shell Tainted: P           O     4.16.2-2-ARCH #1
[    7.468443] Hardware name: ASUS All Series/Z87I-PRO, BIOS 1005 11/26/2014
[    7.468445] RIP: 0010:usercopy_warn+0x7e/0xa0
[    7.468445] RSP: 0018:ffffa6e1c29e7b58 EFLAGS: 00010286
[    7.468446] RAX: 0000000000000000 RBX: ffff9b1636655cb0 RCX: 0000000000000001
[    7.468447] RDX: 0000000080000001 RSI: ffffffff84e67e3c RDI: 00000000ffffffff
[    7.468448] RBP: 0000000000000003 R08: 0000000000000098 R09: 000000000000033b
[    7.468448] R10: ffffffff84ea4121 R11: 0000000000000001 R12: 0000000000000001
[    7.468449] R13: ffff9b1636655cb3 R14: 0000000000000000 R15: ffff9b1636655cf8
[    7.468450] FS:  00007ff53cd92a00(0000) GS:ffff9b165ed00000(0000) knlGS:0000000000000000
[    7.468450] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    7.468451] CR2: 0000559fb185ce38 CR3: 00000003f64ce006 CR4: 00000000001606e0
[    7.468452] Call Trace:
[    7.468455]  __check_object_size+0x130/0x1a0
[    7.468557]  os_memcpy_to_user+0x21/0x40 [nvidia]
[    7.468670]  _nv009377rm+0xbf/0xe0 [nvidia]
[    7.468760]  ? _nv028067rm+0x79/0x90 [nvidia]
[    7.468850]  ? _nv028067rm+0x55/0x90 [nvidia]
[    7.468932]  ? _nv013694rm+0xee/0x100 [nvidia]
[    7.469015]  ? _nv015342rm+0x154/0x270 [nvidia]
[    7.469126]  ? _nv008310rm+0x134/0x1a0 [nvidia]
[    7.469237]  ? _nv008289rm+0x29c/0x2b0 [nvidia]
[    7.469347]  ? _nv001072rm+0xe/0x20 [nvidia]
[    7.469459]  ? _nv007316rm+0xd8/0x100 [nvidia]
[    7.469568]  ? _nv001171rm+0x627/0x830 [nvidia]
[    7.469677]  ? rm_ioctl+0x73/0x100 [nvidia]
[    7.469732]  ? nvidia_ioctl+0x573/0x720 [nvidia]
[    7.469734]  ? kmem_cache_free+0x1bd/0x1f0
[    7.469788]  ? nvidia_frontend_unlocked_ioctl+0x3e/0x50 [nvidia]
[    7.469790]  ? do_vfs_ioctl+0xa4/0x630
[    7.469792]  ? __fput+0x131/0x1e0
[    7.469793]  ? SyS_ioctl+0x74/0x80
[    7.469795]  ? do_syscall_64+0x74/0x190
[    7.469798]  ? entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[    7.469799] Code: 48 c7 c0 e9 20 e7 84 48 0f 44 c2 41 50 51 41 51 48 89 f9 49 89 f1 4d 89 d8 4c 89 d2 48 89 c6 48 c7 c7 40 21 e7 84 e8 02 a7 e3 ff <0f> 0b 48 83 c4 18 c3 48 c7 c6 3c 7e e6 84 49 89 f1 49 89 f3 eb 
[    7.469818] ---[ end trace ffba71c95aa2554a ]---

I am waiting for a new kernel/or nvidia package

Very similar / identical error to mine:

[   18.193149] ------------[ cut here ]------------
[   18.193151] Bad or missing usercopy whitelist? Kernel memory exposure attempt detected from SLUB object 'nvidia_stack_cache' (offset 11440, size 3)!
[   18.193160] WARNING: CPU: 2 PID: 804 at mm/usercopy.c:81 usercopy_warn+0x7e/0xa0
[   18.193160] Modules linked in: bnep nct6775 hwmon_vid snd_hda_codec_hdmi intel_rapl x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm snd_hda_codec_realtek snd_hda_codec_ca0132 snd_hda_codec_generic irqbypass intel_cstate arc4 ofpart cmdlinepart ath9k intel_spi_platform intel_spi spi_nor ath9k_common mtd ath9k_hw iTCO_wdt iTCO_vendor_support wmi_bmof eeepc_wmi fuse ath asus_wmi sparse_keymap mxm_wmi mac80211 ath3k btusb snd_hda_intel btrtl btbcm btintel i915 snd_hda_codec intel_uncore cfg80211 bluetooth snd_hda_core intel_rapl_perf snd_hwdep snd_pcm e1000e joydev mousedev snd_timer mei_me snd ecdh_generic input_leds i2c_algo_bit ptp led_class rfkill i2c_i801 mei lpc_ich rtc_cmos soundcore pps_core shpchp intel_gtt wmi evdev mac_hid nf_log_ipv4 nf_log_common xt_LOG ipt_REJECT nf_reject_ipv4
[   18.193193]  xt_limit xt_tcpudp xt_addrtype nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack ip6_tables nf_conntrack_netbios_ns nf_conntrack_broadcast nf_nat_ftp nf_nat nf_conntrack_ftp nf_conntrack libcrc32c crc32c_generic iptable_filter vboxnetflt(O) vboxnetadp(O) vboxpci(O) vboxdrv(O) sg crypto_user ip_tables x_tables ext4 crc16 mbcache jbd2 fscrypto algif_skcipher af_alg dm_crypt dm_mod hid_generic usbhid hid sr_mod cdrom sd_mod crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel pcbc ahci libahci aesni_intel aes_x86_64 ehci_pci crypto_simd xhci_pci glue_helper libata xhci_hcd ehci_hcd cryptd scsi_mod usbcore usb_common nvidia_drm(PO) drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops drm agpgart nvidia_uvm(PO) nvidia_modeset(PO) nvidia(PO) ipmi_devintf ipmi_msghandler
[   18.193222] CPU: 2 PID: 804 Comm: Xorg Tainted: P           O     4.16.3-1-custom #2
[   18.193223] Hardware name: ASUS All Series/Z87-PRO, BIOS 2103 08/18/2014
[   18.193225] RIP: 0010:usercopy_warn+0x7e/0xa0
[   18.193225] RSP: 0018:ffff9d9d84eebb58 EFLAGS: 00010286
[   18.193226] RAX: 0000000000000000 RBX: ffff99d5d7dcacb0 RCX: 0000000000000001
[   18.193227] RDX: 0000000080000001 RSI: ffffffffbbe67e3c RDI: 00000000ffffffff
[   18.193228] RBP: 0000000000000003 R08: 0000000000000098 R09: 00000000000004fb
[   18.193228] R10: ffffffffbbea4121 R11: 0000000000000001 R12: 0000000000000001
[   18.193229] R13: ffff99d5d7dcacb3 R14: 0000000000000000 R15: ffff99d5d7dcacf8
[   18.193230] FS:  00007fbe1b7f7940(0000) GS:ffff99d5ffa80000(0000) knlGS:0000000000000000
[   18.193231] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   18.193231] CR2: 00007fbe1377d010 CR3: 00000008171de002 CR4: 00000000001606e0
[   18.193232] Call Trace:
[   18.193236]  __check_object_size+0x130/0x1a0
[   18.193348]  os_memcpy_to_user+0x21/0x40 [nvidia]
[   18.193471]  _nv009377rm+0xbf/0xe0 [nvidia]
[   18.193569]  ? _nv028067rm+0x79/0x90 [nvidia]
[   18.193674]  ? _nv028067rm+0x55/0x90 [nvidia]
[   18.193757]  ? _nv013694rm+0xee/0x100 [nvidia]
[   18.193840]  ? _nv015342rm+0x154/0x270 [nvidia]
[   18.193951]  ? _nv008310rm+0x134/0x1a0 [nvidia]
[   18.194062]  ? _nv008289rm+0x29c/0x2b0 [nvidia]
[   18.194172]  ? _nv001072rm+0xe/0x20 [nvidia]
[   18.194284]  ? _nv007316rm+0xd8/0x100 [nvidia]
[   18.194393]  ? _nv001171rm+0x627/0x830 [nvidia]
[   18.194502]  ? rm_ioctl+0x73/0x100 [nvidia]
[   18.194557]  ? nvidia_ioctl+0x573/0x720 [nvidia]
[   18.194559]  ? kmem_cache_free+0x1bd/0x1f0
[   18.194613]  ? nvidia_frontend_unlocked_ioctl+0x3e/0x50 [nvidia]
[   18.194615]  ? do_vfs_ioctl+0xa4/0x630
[   18.194617]  ? __fput+0x131/0x1e0
[   18.194619]  ? preempt_count_add+0x68/0xa0
[   18.194620]  ? SyS_ioctl+0x74/0x80
[   18.194622]  ? do_syscall_64+0x74/0x190
[   18.194625]  ? entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[   18.194626] Code: 48 c7 c0 e9 20 e7 bb 48 0f 44 c2 41 50 51 41 51 48 89 f9 49 89 f1 4d 89 d8 4c 89 d2 48 89 c6 48 c7 c7 40 21 e7 bb e8 b2 a6 e3 ff <0f> 0b 48 83 c4 18 c3 48 c7 c6 3c 7e e6 bb 49 89 f1 49 89 f3 eb 
[   18.194645] ---[ end trace 7438497ac6018632 ]---

---

Please vote for all the AUR packages you're using. You can mass-vote for all of them by doing: "pacman -Qqm | xargs aurvote -v" (make sure to run "aurvote --configure"  first)

brainplot

Member

Registered: 2018-01-27

Posts: 33

Re: Terrible performance regression with Nvidia 390.25 driver

So do you guys confirm that everything is fixed and I can update? I've also been stuck to nvidia 387.34 and linux 4.15

I'm using Chrome 68.0.3409.2 through the dev branch. Do they have things to be fixed on their part in regard to this issue?

phunni

Member

From: Bristol, UK

Registered: 2003-08-13

Posts: 787

Re: Terrible performance regression with Nvidia 390.25 driver

I still have some issues with a few games, although I don't play them often and can't be certain it's the same issue. Most other games seem fine and chromium is also fine.

brainplot

Member

Registered: 2018-01-27

Posts: 33

Re: Terrible performance regression with Nvidia 390.25 driver

I don't play games at all on this machine. I initially had to downgrade because the overall interaction didn't feel very "snappy". KDE animations were a bit jerky and, additionally, Chrome's sandbox had evident problems with the Nvidia driver. (I was using the Stable branch.)

Are these things fixed with the new version? This is mostly a development machine so, like I said, no games are being played here. I'm more interested in the responsiveness of my desktop.

Last edited by brainplot (2018-04-30 09:45:29)

kokoko3k

Member

Registered: 2008-11-14

Posts: 2,421

Re: Terrible performance regression with Nvidia 390.25 driver

Why don't you just try? Works for me.

Last edited by kokoko3k (2018-04-30 20:13:27)

---

Help me to improve ssh-rdp !
Retroarch User? Try my koko-aio shader !

brainplot

Member

Registered: 2018-01-27

Posts: 33

Re: Terrible performance regression with Nvidia 390.25 driver

Why don't you just try? Works for me.

Yeah, that's what I'm going to do. I just didn't want to break my system so I came here to ask, prior to upgrading. That's all.