Kernel 3.18.8 for ROSA 2012 2012.1 2014.1 OMV2013 2014 2015
11 posts
• Page 1 of 1
Kernel 3.18.8 for ROSA 2012 2012.1 2014.1 OMV2013 2014 2015
Kernel 3.18.8 for ROSA 2012 2012.1 2014.1 OMV2013 2014 2015
Kernel 3.18.8 for ROSA 2012lts, rosa2012.1, omv2013.0, omv2014.0, rosa2014.1 and cooker (omv2015.0)
Kernel 3.18.8 - http://lwn.net/Articles/635124/
Our changelog
Build status & Availability
The rpms will be available for for rosa2012lts, rosa 2012.1, omv2013, omv2014, cooker
These below are the rebuilds right now in progress... related to the new kernel 3.18.8
nrj >>> https://abf.rosalinux.ru/kernels_3_18x/ ... uild_lists
- kernel-nrj-desktop
nrjQL >>> https://abf.rosalinux.ru/kernels_3_18x/ ... uild_lists
- kernel-nrjQL-desktop
VMWare virtualization
General Info about kernel 3.18
http://kernelnewbies.org/Linux_3.18
let me know your feedback...
bye, NicCo
Kernel 3.18.8 for ROSA 2012lts, rosa2012.1, omv2013.0, omv2014.0, rosa2014.1 and cooker (omv2015.0)
Kernel 3.18 arrives ASAP in its complex nrj & nrjQL flavours, for ROSA and OpenMandriva Linux OSes!
This is the 8th version of "nrj / nrjQL" stable 3.18.x, in its full development stage,
all features have been already added and completed, it should be a safe update...
Kernel 3.18.8 - http://lwn.net/Articles/635124/
Our changelog
- Code: Select all
%changelog
* Wed Mar 04 2015 Nicolo' Costanza <abitrules@yahoo.it> 3.18.8-1
+ update to 3.18.8 - stable
- drop TOI: there are still issues with hybernation/resume in some HW,
- the four patches for TOI support have been moved to /archives
- add 0001-drm-radeon-dp-Set-EDP_CONFIGURATION_SET-for-bridge-c.patch
- a try to fix : https://issues.openmandriva.org/show_bug.cgi?id=1137
- small fixes and cleanups
- ---------------------------------------------------------------------
- Kernel 3.18 for mdv 2010.2, 2011.0, cooker, rosa.lts2012.0, rosa2012.1
- MIB (Mandriva International Backports) - http://mib.pianetalinux.org/
- The rel (-1) (mainline serie), with official kernel sources and addons,
- the rel (-69) will be used for development and experimental flavours,
- instead (-70) is born by the -1 % -69 merge, can generate all flavours
- Yin & Yang (69) release - it's a very complete kernel flavour sets
- ---------------------------------------------------------------------
* Fri Feb 27 2015 Nicolo' Costanza <abitrules@yahoo.it> 3.18.7-1
+ update to 3.18.7 - stable
- add support for the reiser4 fs with: reiser4-for-3.18.6.patch
- add TOI for 3.18 plus 3 fix patches: TOI is back!
- drop some patches patches that now are already in mainstream
- fix a kconflicts for nvidia-long-lived when mdvver >= 201300
- small fixes and cleanups
- ---------------------------------------------------------------------
- Kernel 3.18 for mdv 2010.2, 2011.0, cooker, rosa.lts2012.0, rosa2012.1
- MIB (Mandriva International Backports) - http://mib.pianetalinux.org/
- The rel (-1) (mainline serie), with official kernel sources and addons,
- the rel (-69) will be used for development and experimental flavours,
- instead (-70) is born by the -1 % -69 merge, can generate all flavours
- Yin & Yang (69) release - it's a very complete kernel flavour sets
- ---------------------------------------------------------------------
* Wed Feb 25 2015 Nicolo' Costanza <abitrules@yahoo.it> 3.18.6-1
+ update to 3.18.6 - stable
- small fixes and cleanups
- ---------------------------------------------------------------------
- Kernel 3.18 for mdv 2010.2, 2011.0, cooker, rosa.lts2012.0, rosa2012.1
- MIB (Mandriva International Backports) - http://mib.pianetalinux.org/
- The rel (-1) (mainline serie), with official kernel sources and addons,
- the rel (-69) will be used for development and experimental flavours,
- instead (-70) is born by the -1 % -69 merge, can generate all flavours
- Yin & Yang (69) release - it's a very complete kernel flavour sets
- ---------------------------------------------------------------------
* Mon Feb 02 2015 Nicolo' Costanza <abitrules@yahoo.it> 3.18.5-1
+ update to 3.18.5 - stable
- drop some patches patches that now are already in mainstream
- small fixes and cleanups
- ---------------------------------------------------------------------
- Kernel 3.18 for mdv 2010.2, 2011.0, cooker, rosa.lts2012.0, rosa2012.1
- MIB (Mandriva International Backports) - http://mib.pianetalinux.org/
- The rel (-1) (mainline serie), with official kernel sources and addons,
- the rel (-69) will be used for development and experimental flavours,
- instead (-70) is born by the -1 % -69 merge, can generate all flavours
- Yin & Yang (69) release - it's a very complete kernel flavour sets
- ---------------------------------------------------------------------
* Sat Jan 31 2015 Nicolo' Costanza <abitrules@yahoo.it> 3.18.4-1
+ update to 3.18.4 - stable
- add a new key > Fibre Channel over Ethernet (FCoE) (I40E_FCOE) [N/y/?] (NEW)
- currently its first setup value as disabled > # CONFIG_I40E_FCOE is not set
- drop some patches patches that now are already in mainstream
- small fixes and cleanups
- ---------------------------------------------------------------------
- Kernel 3.18 for mdv 2010.2, 2011.0, cooker, rosa.lts2012.0, rosa2012.1
- MIB (Mandriva International Backports) - http://mib.pianetalinux.org/
- The rel (-1) (mainline serie), with official kernel sources and addons,
- the rel (-69) will be used for development and experimental flavours,
- instead (-70) is born by the -1 % -69 merge, can generate all flavours
- Yin & Yang (69) release - it's a very complete kernel flavour sets
- ---------------------------------------------------------------------
* Tue Jan 27 2015 Nicolo' Costanza <abitrules@yahoo.it> 3.18.3-1
+ update to 3.18.3 - stable
- drop # block fixes patches that now are already in mainstream
- small fixes and cleanups
- ---------------------------------------------------------------------
- Kernel 3.18 for mdv 2010.2, 2011.0, cooker, rosa.lts2012.0, rosa2012.1
- MIB (Mandriva International Backports) - http://mib.pianetalinux.org/
- The rel (-1) (mainline serie), with official kernel sources and addons,
- the rel (-69) will be used for development and experimental flavours,
- instead (-70) is born by the -1 % -69 merge, can generate all flavours
- Yin & Yang (69) release - it's a very complete kernel flavour sets
- ---------------------------------------------------------------------
* Fri Jan 23 2015 Nicolo' Costanza <abitrules@yahoo.it> 3.18.2-1
+ update to 3.18.2 - stable
- add a patch to fix performance issue:
- x86-vdso-Use-asm-volatile-in-__getcpu.patch
- add a patch for iwlwifi:
- net-wireless-iwlwifi-mvm-fix-Rx-with-both-chains.patch
- ---------------------------------------------------------------------
- Kernel 3.18 for mdv 2010.2, 2011.0, cooker, rosa.lts2012.0, rosa2012.1
- MIB (Mandriva International Backports) - http://mib.pianetalinux.org/
- The rel (-1) (mainline serie), with official kernel sources and addons,
- the rel (-69) will be used for development and experimental flavours,
- instead (-70) is born by the -1 % -69 merge, can generate all flavours
- Yin & Yang (69) release - it's a very complete kernel flavour sets
- ---------------------------------------------------------------------
* Wed Jan 21 2015 Nicolo' Costanza <abitrules@yahoo.it> 3.18.1-1
+ update to 3.18.1 - stable
+ it's first version of "nrj/nrjQL" stable 3.18.x, in early development
- stage, so, it's only for testing purposes, please, dont use this srpm,
- for your daily PC working, because is still to fix all over!!!
- all the defconfigs have been prepared for the 3.18 series
- all the patches have been added/update for the 3.18 series
- all the kernel specs have been updated to the 3.18 series
- all custom openmandriva/rosa features have been backported from previous
- ---------------------------------------------------------------------
- Kernel 3.18 for mdv 2010.2, 2011.0, cooker, rosa.lts2012.0, rosa2012.1
- MIB (Mandriva International Backports) - http://mib.pianetalinux.org/
- The rel (-1) (mainline serie), with official kernel sources and addons,
- the rel (-69) will be used for development and experimental flavours,
- instead (-70) is born by the -1 % -69 merge, can generate all flavours
- Yin & Yang (69) release - it's a very complete kernel flavour sets
- ---------------------------------------------------------------------
* Mon Jan 12 2015 Nicolo' Costanza <abitrules@yahoo.it> 3.17.8-1
+ update to 3.17.8 - stable
- ---------------------------------------------------------------------
- Kernel 3.17 for mdv 2010.2, 2011.0, cooker, rosa.lts2012.0, rosa2012.1
- MIB (Mandriva International Backports) - http://mib.pianetalinux.org/
- The rel (-1) (mainline serie), with official kernel sources and addons,
- the rel (-QL) will be used for development and experimental flavours,
- instead (-ONE) is born by the -1 % -69 merge, can generate all flavours
- Yin & Yang (69) release - it's a very complete kernel flavour sets
- ---------------------------------------------------------------------
Build status & Availability
I remind you the base repo address for kernels 3.18 link:
http://abf.rosalinux.ru/downloads/kerne ... epository/
The rpms will be available for for rosa2012lts, rosa 2012.1, omv2013, omv2014, cooker
These below are the rebuilds right now in progress... related to the new kernel 3.18.8
nrj >>> https://abf.rosalinux.ru/kernels_3_18x/ ... uild_lists
- kernel-nrj-desktop
nrjQL >>> https://abf.rosalinux.ru/kernels_3_18x/ ... uild_lists
- kernel-nrjQL-desktop
VMWare virtualization
NEWS:
With current VMware 11 / VMPlayer 7: everything works without requiring any unofficial patching
General Info about kernel 3.18
http://kernelnewbies.org/Linux_3.18
let me know your feedback...
bye, NicCo
.
--- Professional experience ---
Kernel designer, engineer, maintainer and tester for ROSA Desktop and OpenMandriva Lx O.S.
--- currently I'm playing with ---
LTS Kernels > Linux 4.1.12-nrjQL <<< Linux 3.18.17-nrjQL <<< Linux 3.14.46-nrjQL
EOL Kernels > Linux 3.19.8-nrjQL <<< Linux 3.17.8-nrjQL <<< Linux 3.15.10-nrjQL
--- Professional experience ---
Kernel designer, engineer, maintainer and tester for ROSA Desktop and OpenMandriva Lx O.S.
--- currently I'm playing with ---
LTS Kernels > Linux 4.1.12-nrjQL <<< Linux 3.18.17-nrjQL <<< Linux 3.14.46-nrjQL
EOL Kernels > Linux 3.19.8-nrjQL <<< Linux 3.17.8-nrjQL <<< Linux 3.15.10-nrjQL
-
NicCo
- Amministratore
- Posts: 4737
- Joined: 6 December 2007, 19:54
Hi!
With this version i dont have the black screen after suspend on ram!
Good!
With this version i dont have the black screen after suspend on ram!
Good!
OMA-2014
oiram/bin/selom
Da ognuno secondo le proprie capacità ad ognuno secondo i propri bisogni!
oiram/bin/selom
Da ognuno secondo le proprie capacità ad ognuno secondo i propri bisogni!
-
oiram73
- Utente senior
- Posts: 405
- Joined: 5 January 2009, 13:33
- Location: Perugia
- ROSA: -
- OpenMandriva: 2014
- Kernel: 3.19.8-nrjQL-realtim
- Desktop: KDE 4.14.3
- country: Italia
oiram73 wrote:Hi!
With this version i dont have the black screen after suspend on ram!
Good!
Eureka! Good!
So, the proposed patch really fix it!
It was a very long time nasty issue,,,.
from changelog
- add 0001-drm-radeon-dp-Set-EDP_CONFIGURATION_SET-for-bridge-c.patch
- a try to fix : https://issues.openmandriva.org/show_bug.cgi?id=1137
DId you had a different bugzilla open issue?
If so, tell me the number to be closed as solved
Mario Moles wrote:Solo che con gli ATI-open NON si riaccende lo schermo cosa che ho già segnalato qui:
https://issues.openmandriva.org/show_bug.cgi?id=732
solo che finora non ho ancora avuto modo di insistere.
.
--- Professional experience ---
Kernel designer, engineer, maintainer and tester for ROSA Desktop and OpenMandriva Lx O.S.
--- currently I'm playing with ---
LTS Kernels > Linux 4.1.12-nrjQL <<< Linux 3.18.17-nrjQL <<< Linux 3.14.46-nrjQL
EOL Kernels > Linux 3.19.8-nrjQL <<< Linux 3.17.8-nrjQL <<< Linux 3.15.10-nrjQL
--- Professional experience ---
Kernel designer, engineer, maintainer and tester for ROSA Desktop and OpenMandriva Lx O.S.
--- currently I'm playing with ---
LTS Kernels > Linux 4.1.12-nrjQL <<< Linux 3.18.17-nrjQL <<< Linux 3.14.46-nrjQL
EOL Kernels > Linux 3.19.8-nrjQL <<< Linux 3.17.8-nrjQL <<< Linux 3.15.10-nrjQL
-
NicCo
- Amministratore
- Posts: 4737
- Joined: 6 December 2007, 19:54
OMA-2014
oiram/bin/selom
Da ognuno secondo le proprie capacità ad ognuno secondo i propri bisogni!
oiram/bin/selom
Da ognuno secondo le proprie capacità ad ognuno secondo i propri bisogni!
-
oiram73
- Utente senior
- Posts: 405
- Joined: 5 January 2009, 13:33
- Location: Perugia
- ROSA: -
- OpenMandriva: 2014
- Kernel: 3.19.8-nrjQL-realtim
- Desktop: KDE 4.14.3
- country: Italia
Now it performs much better than 3.18.7. My PC suspends to RAM correctly, only problem I have is with no network after resume. I can not start network.service, here is output:
Only reboot helps to bring network up.
- Code: Select all
[root@localhost ~]# systemctl status network.service
network.service - LSB: Bring up/down networking
Loaded: loaded (/etc/rc.d/init.d/network)
Active: failed (Result: exit-code) since Sat 2015-03-07 08:17:10 CET; 35s ago
Process: 10490 ExecStart=/etc/rc.d/init.d/network start (code=exited, status=1/FAILURE)
Mar 07 08:17:06 localhost.localdomain network[10490]: RTNETLINK answers: File exists
Mar 07 08:17:07 localhost.localdomain network[10490]: RTNETLINK answers: File exists
Mar 07 08:17:08 localhost.localdomain network[10490]: RTNETLINK answers: File exists
Mar 07 08:17:08 localhost.localdomain network[10490]: RTNETLINK answers: File exists
Mar 07 08:17:09 localhost.localdomain network[10490]: RTNETLINK answers: File exists
Mar 07 08:17:10 localhost.localdomain network[10490]: RTNETLINK answers: File exists
Mar 07 08:17:10 localhost.localdomain network[10490]: RTNETLINK answers: File exists
Mar 07 08:17:10 localhost.localdomain systemd[1]: network.service: control process exited, code=exited status=1
Mar 07 08:17:10 localhost.localdomain systemd[1]: Failed to start LSB: Bring up/down networking.
Mar 07 08:17:10 localhost.localdomain systemd[1]: Unit network.service entered failed state.
root@localhost ~]# journalctl -xn
-- Logs begin at Sat 2015-02-21 21:21:26 CET, end at Sat 2015-03-07 08:19:01 CET. --
Mar 07 08:18:59 localhost.localdomain kernel: ieee80211 phy0: rt61pci_wait_bbp_ready: Error - BBP register access failed, aborting
Mar 07 08:18:59 localhost.localdomain kernel: ieee80211 phy0: rt61pci_set_device_state: Error - Device failed to enter state 4 (-5)
Mar 07 08:19:00 localhost.localdomain kernel: ieee80211 phy0: rt61pci_wait_bbp_ready: Error - BBP register access failed, aborting
Mar 07 08:19:00 localhost.localdomain kernel: ieee80211 phy0: rt61pci_set_device_state: Error - Device failed to enter state 4 (-5)
Mar 07 08:19:01 localhost.localdomain kernel: ieee80211 phy0: rt61pci_wait_bbp_ready: Error - BBP register access failed, aborting
Mar 07 08:19:01 localhost.localdomain kernel: ieee80211 phy0: rt61pci_set_device_state: Error - Device failed to enter state 4 (-5)
Mar 07 08:19:01 localhost.localdomain NetworkManager[3955]: <error> [1425712741.671032] [nm-supplicant-interface.c:997] interface_add_cb(): (wlan0): error adding interface: wpa_supplicant couldn't grab this inte
Mar 07 08:19:01 localhost.localdomain NetworkManager[3955]: dbus_g_proxy_cancel_call: assertion 'pending != NULL' failed
Mar 07 08:19:01 localhost.localdomain NetworkManager[3955]: <info> (wlan0): supplicant interface state: starting -> down
Mar 07 08:19:01 localhost.localdomain NetworkManager[3955]: <warn> Trying to remove a non-existant call id.
04:00.0 Network controller: Ralink corp. RT2561/RT61 802.11g PCI
Subsystem: Belkin F5D7000 v6000 Wireless G Desktop Card
Control: I/O- Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=slow >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
Interrupt: pin A routed to IRQ 19
Region 0: Memory at faff8000 (32-bit, non-prefetchable) [size=32K]
Capabilities: [40] Power Management version 2
Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
Kernel driver in use: rt61pci
Only reboot helps to bring network up.
-
trencan
- Utente junior
- Posts: 24
- Joined: 27 April 2013, 16:33
- ROSA: 2012lts
- OpenMandriva: -
- Kernel: 3.8.8
- Desktop: KDE
- country: Czech
Installed 3.18.8-nrjQL on all my machines: all works good! 

Desktop A: Kubuntu 32bit
Desktop B: Kubuntu 64bit
Notebook Asus X53SD-SX191V: Kubuntu 64bit
Linux 4 Ever
Desktop B: Kubuntu 64bit
Notebook Asus X53SD-SX191V: Kubuntu 64bit
Linux 4 Ever
-
Alexander83
- Utente
- Posts: 248
- Joined: 10 April 2012, 14:50
- ROSA: -
- OpenMandriva: -
- Kernel: -
- Desktop: Kde
- country: Italy
Hi trencan, try unload and reload the module of the network adapter.
Bye.
Bye.
Astragalo
-
astragalo
- Collaboratore
- Posts: 641
- Joined: 7 December 2007, 18:29
- OpenMandriva: 2010.2
- Kernel: 2.6.36.3-nrj-MIB
- Desktop: Kde 4.5.5
astragalo wrote:Hi trencan, try unload and reload the module of the network adapter.
Bye.
I tried that, it didn't help, here is result:
- Code: Select all
root@localhost ~]# systemctl restart network.service
Job for network.service failed. See 'systemctl status network.service' and 'journalctl -xn' for details.
[root@localhost ~]# systemctl status network.service
network.service - LSB: Bring up/down networking
Loaded: loaded (/etc/rc.d/init.d/network)
Active: failed (Result: exit-code) since Mon 2015-03-09 18:19:41 CET; 9s ago
Process: 15191 ExecStart=/etc/rc.d/init.d/network start (code=exited, status=1/FAILURE)
Mar 09 18:19:37 localhost.localdomain network[15191]: RTNETLINK answers: File exists
Mar 09 18:19:38 localhost.localdomain network[15191]: RTNETLINK answers: File exists
Mar 09 18:19:39 localhost.localdomain network[15191]: RTNETLINK answers: File exists
Mar 09 18:19:39 localhost.localdomain network[15191]: RTNETLINK answers: File exists
Mar 09 18:19:40 localhost.localdomain network[15191]: RTNETLINK answers: File exists
Mar 09 18:19:41 localhost.localdomain network[15191]: RTNETLINK answers: File exists
Mar 09 18:19:41 localhost.localdomain network[15191]: RTNETLINK answers: File exists
Mar 09 18:19:41 localhost.localdomain systemd[1]: network.service: control process exited, code=exited status=1
Mar 09 18:19:41 localhost.localdomain systemd[1]: Failed to start LSB: Bring up/down networking.
Mar 09 18:19:41 localhost.localdomain systemd[1]: Unit network.service entered failed state.
[root@localhost ~]# modprobe -r rt61pci
[root@localhost ~]# lsmod | grep rt61pci
[root@localhost ~]# modprobe rt61pci
[root@localhost ~]# lsmod | grep rt61pci
rt61pci 32088 0
rt2x00pci 13287 1 rt61pci
rt2x00mmio 13603 1 rt61pci
rt2x00lib 55250 3 rt61pci,rt2x00pci,rt2x00mmio
eeprom_93cx6 13344 1 rt61pci
crc_itu_t 12707 1 rt61pci
[root@localhost ~]# systemctl restart network.service
Job for network.service failed. See 'systemctl status network.service' and 'journalctl -xn' for details.
[root@localhost ~]# journalctl -xn
-- Logs begin at Tue 2015-02-24 05:53:28 CET, end at Mon 2015-03-09 18:21:30 CET. --
Mar 09 18:21:28 localhost.localdomain kernel: ieee80211 phy0: rt61pci_wait_bbp_ready: Error - BBP register access failed, aborting
Mar 09 18:21:28 localhost.localdomain kernel: ieee80211 phy0: rt61pci_set_device_state: Error - Device failed to enter state 4 (-5)
Mar 09 18:21:29 localhost.localdomain kernel: ieee80211 phy0: rt61pci_wait_bbp_ready: Error - BBP register access failed, aborting
Mar 09 18:21:29 localhost.localdomain kernel: ieee80211 phy0: rt61pci_set_device_state: Error - Device failed to enter state 4 (-5)
Mar 09 18:21:30 localhost.localdomain NetworkManager[3907]: <error> [1425921690.529048] [nm-supplicant-interface.c:997] interface_add_cb(): (wlan0): error adding interface: wpa_supplicant couldn't grab this inte
Mar 09 18:21:30 localhost.localdomain NetworkManager[3907]: dbus_g_proxy_cancel_call: assertion 'pending != NULL' failed
Mar 09 18:21:30 localhost.localdomain NetworkManager[3907]: <info> (wlan0): supplicant interface state: starting -> down
Mar 09 18:21:30 localhost.localdomain NetworkManager[3907]: <warn> Trying to remove a non-existant call id.
Mar 09 18:21:30 localhost.localdomain kernel: ieee80211 phy0: rt61pci_wait_bbp_ready: Error - BBP register access failed, aborting
Mar 09 18:21:30 localhost.localdomain kernel: ieee80211 phy0: rt61pci_set_device_state: Error - Device failed to enter state 4 (-5)
But also if it helps, it is not permanent solution for me, I can not reload kernel module after every resume.
-
trencan
- Utente junior
- Posts: 24
- Joined: 27 April 2013, 16:33
- ROSA: 2012lts
- OpenMandriva: -
- Kernel: 3.8.8
- Desktop: KDE
- country: Czech
Ciao Nicco,
ho provato il 3.18.8, e sembra OK; ho trovato la causa del
problema dello scheduler, ovvero il fatto che il /sys/block/sda/queue/scheduler
è sempre settato a "none". Sembra che con il 3.18 ufficiale sia stato introdotto
una specie di nuovo scheduler, chiamato "Multiblock Queue I/O",
che dovrebbe abbassare la latenza per dischi SSD, quando ci sono tantissime
operazioni (superiori a 1 milione di OP al secondo). Qui c'e' una spiegazione:
http://www.phoronix.com/scan.php?page=n ... px=MTgzMjM
Come puoi leggere al fondo c'e' scritto:
"missing features right now are I/O scheduler support in blk-mq and mailine multi-path support."
Ovvero manca ancora il supporto per gli I/O scheduler.
Infatti se vai su /sys/block/sda/, noterai che c'e' una directory /sys/block/sda/mq/ che
prima non c'era.
A triggherare questo nuovo scheduler è l'opzione CONFIG_SCSI_MQ_DEFAULT, che
nel kernel 3.18.x di OMV/Rosa si è tirato dentro il valore CONFIG_SCSI_MQ_DEFAULT=y. Riportandola
a "# CONFIG_SCSI_MQ_DEFAULT is not set", le cose vengono riportate come prima
e il BFQ, come pure gli altri scheduler riaffiorano.
Ho provato a fare un build veloce con quella chiave disabilitata del 3.18.8 aggiornandolo al 3.18.9, e l'ho messo qui:
http://abf-downloads.rosalinux.ru/ghibo ... n/release/
http://abf-downloads.rosalinux.ru/ghibo ... n/release/
in cui ho disabilitato quell'opzione e il tutto sembra essere tornato ai valori pre Multiblock-queue (avevo
riabililitato il TOI).
Ho sentito anche il Valente poco fa, e secondo lui il multiblock dovrebbe abbassare la latenza e quindi
rendere superfluo lo scheduler. Tuttavia sembra che al momento questo sia falso e l'ha discusso anche al LinuxCon.
Il multiblock dovrebbe dare prestazioni migliori sopra il milione di OPS per dispositivi high-end, dove l'assenza
del multiblock rappresenterebbe un collo di bottiglia. Tuttavia, a detta di Paolo, sembra che al momento (magari miglioreranno in futuro),
le garanzie del multibock siano peggiori di quelle del deadline e in termini di latenza c'e' un bel disastro
(non mi stupirei che fosse anche la causa dei problemi dei trasferimenti USB di cui qui https://forums.openmandriva.org/en/disc ... 8-5#latest).
Ciao.
J.
ho provato il 3.18.8, e sembra OK; ho trovato la causa del
problema dello scheduler, ovvero il fatto che il /sys/block/sda/queue/scheduler
è sempre settato a "none". Sembra che con il 3.18 ufficiale sia stato introdotto
una specie di nuovo scheduler, chiamato "Multiblock Queue I/O",
che dovrebbe abbassare la latenza per dischi SSD, quando ci sono tantissime
operazioni (superiori a 1 milione di OP al secondo). Qui c'e' una spiegazione:
http://www.phoronix.com/scan.php?page=n ... px=MTgzMjM
Come puoi leggere al fondo c'e' scritto:
"missing features right now are I/O scheduler support in blk-mq and mailine multi-path support."
Ovvero manca ancora il supporto per gli I/O scheduler.
Infatti se vai su /sys/block/sda/, noterai che c'e' una directory /sys/block/sda/mq/ che
prima non c'era.
A triggherare questo nuovo scheduler è l'opzione CONFIG_SCSI_MQ_DEFAULT, che
nel kernel 3.18.x di OMV/Rosa si è tirato dentro il valore CONFIG_SCSI_MQ_DEFAULT=y. Riportandola
a "# CONFIG_SCSI_MQ_DEFAULT is not set", le cose vengono riportate come prima
e il BFQ, come pure gli altri scheduler riaffiorano.
Ho provato a fare un build veloce con quella chiave disabilitata del 3.18.8 aggiornandolo al 3.18.9, e l'ho messo qui:
http://abf-downloads.rosalinux.ru/ghibo ... n/release/
http://abf-downloads.rosalinux.ru/ghibo ... n/release/
in cui ho disabilitato quell'opzione e il tutto sembra essere tornato ai valori pre Multiblock-queue (avevo
riabililitato il TOI).
Ho sentito anche il Valente poco fa, e secondo lui il multiblock dovrebbe abbassare la latenza e quindi
rendere superfluo lo scheduler. Tuttavia sembra che al momento questo sia falso e l'ha discusso anche al LinuxCon.
Il multiblock dovrebbe dare prestazioni migliori sopra il milione di OPS per dispositivi high-end, dove l'assenza
del multiblock rappresenterebbe un collo di bottiglia. Tuttavia, a detta di Paolo, sembra che al momento (magari miglioreranno in futuro),
le garanzie del multibock siano peggiori di quelle del deadline e in termini di latenza c'e' un bel disastro
(non mi stupirei che fosse anche la causa dei problemi dei trasferimenti USB di cui qui https://forums.openmandriva.org/en/disc ... 8-5#latest).
Ciao.
J.
-
joeghi
- Nuovo utente
- Posts: 5
- Joined: 27 September 2014, 16:46
- ROSA: -
- OpenMandriva: 2010.2
- Kernel: 3.15.10-nrjQL-realti
- Desktop: KDE
- country: Italy
joeghi wrote:Ciao Nicco,
ho provato il 3.18.8, e sembra OK; ho trovato la causa del
problema dello scheduler, ovvero il fatto che il /sys/block/sda/queue/scheduler
è sempre settato a "none". Sembra che con il 3.18 ufficiale sia stato introdotto
una specie di nuovo scheduler, chiamato "Multiblock Queue I/O",
che dovrebbe abbassare la latenza per dischi SSD, quando ci sono tantissime
operazioni (superiori a 1 milione di OP al secondo). Qui c'e' una spiegazione:
http://www.phoronix.com/scan.php?page=n ... px=MTgzMjM
.........................................................................
Avevo da poco trovato e letto qualcosa circa il nuovo scheduler chiamato "Multiblock Queue I/O"
cambierò la chiave nel modo da te suggerito
# CONFIG_SCSI_MQ_DEFAULT is not set"
sperando questo possa risolvere anche i crash riscontrati durante i trasfer i/0 pesanti su usb
Per quanto riguarda il TOI, io l'avevo disattivato in quanto alcuni avevano riscontrato malfunzionamenti soprattutto durante le operazioni di hybernation o del suo successivo resume, fatto che io non ho potuto riprodurre in quanto sui miei portatili non avevo riscontrato alcun problema, a me funzionava bene tutto del TOI, ma leggendo le numerose lamentele al riguardo, avevo deciso momentaneamente di toglierlo per evitare questi problemi...
grazie,
ciao
.
--- Professional experience ---
Kernel designer, engineer, maintainer and tester for ROSA Desktop and OpenMandriva Lx O.S.
--- currently I'm playing with ---
LTS Kernels > Linux 4.1.12-nrjQL <<< Linux 3.18.17-nrjQL <<< Linux 3.14.46-nrjQL
EOL Kernels > Linux 3.19.8-nrjQL <<< Linux 3.17.8-nrjQL <<< Linux 3.15.10-nrjQL
--- Professional experience ---
Kernel designer, engineer, maintainer and tester for ROSA Desktop and OpenMandriva Lx O.S.
--- currently I'm playing with ---
LTS Kernels > Linux 4.1.12-nrjQL <<< Linux 3.18.17-nrjQL <<< Linux 3.14.46-nrjQL
EOL Kernels > Linux 3.19.8-nrjQL <<< Linux 3.17.8-nrjQL <<< Linux 3.15.10-nrjQL
-
NicCo
- Amministratore
- Posts: 4737
- Joined: 6 December 2007, 19:54
A very good News for the Kernel 3.18 serie, with 3.18.9 it has been declared as LTS, that was really expected / hoped of me...
Linux Kernel 3.18.9 Is Now an LTS (Long-Term Support) Release
http://news.softpedia.com/news/Linux-Ke ... 5556.shtml
Kernel 3.18.9 will follow soon here... and many other following versions...
Linux Kernel 3.18.9 Is Now an LTS (Long-Term Support) Release
http://news.softpedia.com/news/Linux-Ke ... 5556.shtml
Kernel 3.18.9 will follow soon here... and many other following versions...
.
--- Professional experience ---
Kernel designer, engineer, maintainer and tester for ROSA Desktop and OpenMandriva Lx O.S.
--- currently I'm playing with ---
LTS Kernels > Linux 4.1.12-nrjQL <<< Linux 3.18.17-nrjQL <<< Linux 3.14.46-nrjQL
EOL Kernels > Linux 3.19.8-nrjQL <<< Linux 3.17.8-nrjQL <<< Linux 3.15.10-nrjQL
--- Professional experience ---
Kernel designer, engineer, maintainer and tester for ROSA Desktop and OpenMandriva Lx O.S.
--- currently I'm playing with ---
LTS Kernels > Linux 4.1.12-nrjQL <<< Linux 3.18.17-nrjQL <<< Linux 3.14.46-nrjQL
EOL Kernels > Linux 3.19.8-nrjQL <<< Linux 3.17.8-nrjQL <<< Linux 3.15.10-nrjQL
-
NicCo
- Amministratore
- Posts: 4737
- Joined: 6 December 2007, 19:54
11 posts
• Page 1 of 1
Return to Latest from packagers & testing
Who is online
Users browsing this forum: No registered users and 135 guests