So is this a working system or not? Need to see/compare results.
This IS a âworking systemâ . . . what makes the difference is the 6.13xx kernels either donât revive the display OR do so intermittently, sometimes waking the display, sometimes not. Right now I have booted the working consistently 6.12.6 . . . . Would these params be changing within the same install of TW if I changed the kernel to the inconsistent one?
Or, I donât have 6.12.6 installed in my new TW partition, I believe all it has now is 6.13.6 . . . which has also been âinconsistent.â
Otherwise, both systems âworkâ fine, right up until I suspend the machine, then we go into the black and donât come out. My search is to try to have a display that revives after suspend on a consistent basis.
The question is, will these params change if I switch the kernel, in the âoldâ version of TW?? I was assuming that these data are for the install, irregardless of kernel??? Iâd have to get back to you on the data from the brand new TW install.
@non_space I have no idea if they change, but if they do, then that may be part of the problem, I donât use nouveau, so can only surmise based on the information feed, if no information, then probably easier for you to rock on with what works and donât update the system to avoid more issues, sans any securityâŚ
OK . . . well, I can obviously run the basic command with each of the 3 kernels I now have in the Old TW install and then run it in the new . . . .
I have locked in the 6.12.6 kernel so it should be there, and that does seem to make a difference.
The question on the âwashing hands of old hardwareâ gambit is that with the newer kernels, some of them âworkâ intermittently to provide the display after reviving and some donât . . . .
Thus tossing the cards back up into the air again with each new kernel.
Same install running 6.13.5:
# systool -vm nouveau
Module = "nouveau"
Attributes:
coresize = "3641344"
initsize = "0"
initstate = "live"
refcnt = "5"
srcversion = "D7F129F9B8E111A71198A04"
taint = ""
uevent = <store method only>
Parameters:
NVreg_RegistryDwords= "(null)"
atomic = "0"
config = 28 6e 75 6c 6c 29 0a
debug = "(null)"
duallink = "1"
hdmimhz = "0"
ignorelid = "0"
kms_vram_pushbuf = "-1"
modeset = "-1"
mst = "1"
noaccel = "0"
runpm = "-1"
tv_disable = "0"
tv_norm = "(null)"
vram_pushbuf = "0"
Sections:
.altinstr_replacement= "0xffffffffc074539a"
.altinstructions = "0xffffffffc0830000"
.bss = "0xffffffffc0770e80"
.call_sites = "0xffffffffc0906dd0"
.data..once = "0xffffffffc07706d8"
.data..read_mostly = "0xffffffffc0770720"
.data = "0xffffffffc0751b60"
.exit.data = "0xffffffffc07706e0"
.exit.text = "0xffffffffc07453e0"
.gnu.linkonce.this_module= "0xffffffffc0770900"
.ibt_endbr_seal = "0xffffffffc091ce58"
.init.data = "0xffffffffc0983000"
.init.text = "0xffffffffc03a0000"
.note.Linux = "0xffffffffc09816e4"
.note.gnu.build-id = "0xffffffffc09816c0"
.note.gnu.property = "0xffffffffc0981680"
.orc_header = "0xffffffffc0981714"
.orc_unwind = "0xffffffffc091df68"
.orc_unwind_ip = "0xffffffffc0959a10"
.printk_index = "0xffffffffc076a6a8"
.retpoline_sites = "0xffffffffc0900160"
.return_sites = "0xffffffffc09033a0"
.rodata = "0xffffffffc08300c0"
.rodata.cst2 = "0xffffffffc08e89ee"
.rodata.str1.1 = "0xffffffffc08d9ff8"
.rodata.str1.8 = "0xffffffffc08e8a00"
.smp_locks = "0xffffffffc08ffd38"
.static_call_sites = "0xffffffffc0770798"
.strtab = "0xffffffffc09f59b0"
.symtab = "0xffffffffc0983008"
.text = "0xffffffffc0608000"
.text.unlikely = "0xffffffffc0730300"
__bug_table = "0xffffffffc0747000"
__dyndbg = "0xffffffffc0770728"
__dyndbg_classes = "0xffffffffc07706e8"
__jump_table = "0xffffffffc039e000"
__mcount_loc = "0xffffffffc08d1c90"
__param = "0xffffffffc08fff08"
__patchable_function_entries= "0xffffffffc0749788"
Same old install, now running 6.13.6:
# systool -vm nouveau
Module = "nouveau"
Attributes:
coresize = "3641344"
initsize = "0"
initstate = "live"
refcnt = "5"
srcversion = "D7F129F9B8E111A71198A04"
taint = ""
uevent = <store method only>
Parameters:
NVreg_RegistryDwords= "(null)"
atomic = "0"
config = 28 6e 75 6c 6c 29 0a
debug = "(null)"
duallink = "1"
hdmimhz = "0"
ignorelid = "0"
kms_vram_pushbuf = "-1"
modeset = "-1"
mst = "1"
noaccel = "0"
runpm = "-1"
tv_disable = "0"
tv_norm = "(null)"
vram_pushbuf = "0"
Sections:
.altinstr_replacement= "0xffffffffc071d39a"
.altinstructions = "0xffffffffc0808000"
.bss = "0xffffffffc0748e80"
.call_sites = "0xffffffffc08dedd0"
.data..once = "0xffffffffc07486d8"
.data..read_mostly = "0xffffffffc0748720"
.data = "0xffffffffc0729b60"
.exit.data = "0xffffffffc07486e0"
.exit.text = "0xffffffffc071d3e0"
.gnu.linkonce.this_module= "0xffffffffc0748900"
.ibt_endbr_seal = "0xffffffffc08f4e58"
.init.data = "0xffffffffc095b000"
.init.text = "0xffffffffc0487000"
.note.Linux = "0xffffffffc09596e4"
.note.gnu.build-id = "0xffffffffc09596c0"
.note.gnu.property = "0xffffffffc0959680"
.orc_header = "0xffffffffc0959714"
.orc_unwind = "0xffffffffc08f5f68"
.orc_unwind_ip = "0xffffffffc0931a10"
.printk_index = "0xffffffffc07426a8"
.retpoline_sites = "0xffffffffc08d8160"
.return_sites = "0xffffffffc08db3a0"
.rodata = "0xffffffffc08080c0"
.rodata.cst2 = "0xffffffffc08c09ee"
.rodata.str1.1 = "0xffffffffc08b1ff8"
.rodata.str1.8 = "0xffffffffc08c0a00"
.smp_locks = "0xffffffffc08d7d38"
.static_call_sites = "0xffffffffc0748798"
.strtab = "0xffffffffc09cd9b0"
.symtab = "0xffffffffc095b008"
.text = "0xffffffffc05e0000"
.text.unlikely = "0xffffffffc0708300"
__bug_table = "0xffffffffc071f000"
__dyndbg = "0xffffffffc0748728"
__dyndbg_classes = "0xffffffffc07486e8"
__jump_table = "0xffffffffc047e000"
__mcount_loc = "0xffffffffc08a9c90"
__param = "0xffffffffc08d7f08"
__patchable_function_entries= "0xffffffffc0721788"
Command run on the fresh install of TW, running 6.13.6 and which has âNvidia serviceâ repo enabled, but running nouveau:
# systool -vm nouveau
Module = "nouveau"
Attributes:
coresize = "3641344"
initsize = "0"
initstate = "live"
refcnt = "10"
srcversion = "D7F129F9B8E111A71198A04"
taint = ""
uevent = <store method only>
Parameters:
NVreg_RegistryDwords= "(null)"
atomic = "0"
config = 28 6e 75 6c 6c 29 0a
debug = "(null)"
duallink = "1"
hdmimhz = "0"
ignorelid = "0"
kms_vram_pushbuf = "-1"
modeset = "-1"
mst = "1"
noaccel = "0"
runpm = "-1"
tv_disable = "0"
tv_norm = "(null)"
vram_pushbuf = "0"
Sections:
.altinstr_replacement= "0xffffffffc076a39a"
.altinstructions = "0xffffffffc0855000"
.bss = "0xffffffffc0795e80"
.call_sites = "0xffffffffc092bdd0"
.data..once = "0xffffffffc07956d8"
.data..read_mostly = "0xffffffffc0795720"
.data = "0xffffffffc0776b60"
.exit.data = "0xffffffffc07956e0"
.exit.text = "0xffffffffc076a3e0"
.gnu.linkonce.this_module= "0xffffffffc0795900"
.ibt_endbr_seal = "0xffffffffc0941e58"
.init.data = "0xffffffffc09a8000"
.init.text = "0xffffffffc03f1000"
.note.Linux = "0xffffffffc09a66e4"
.note.gnu.build-id = "0xffffffffc09a66c0"
.note.gnu.property = "0xffffffffc09a6680"
.orc_header = "0xffffffffc09a6714"
.orc_unwind = "0xffffffffc0942f68"
.orc_unwind_ip = "0xffffffffc097ea10"
.printk_index = "0xffffffffc078f6a8"
.retpoline_sites = "0xffffffffc0925160"
.return_sites = "0xffffffffc09283a0"
.rodata = "0xffffffffc08550c0"
.rodata.cst2 = "0xffffffffc090d9ee"
.rodata.str1.1 = "0xffffffffc08feff8"
.rodata.str1.8 = "0xffffffffc090da00"
.smp_locks = "0xffffffffc0924d38"
.static_call_sites = "0xffffffffc0795798"
.strtab = "0xffffffffc0a1a9b0"
.symtab = "0xffffffffc09a8008"
.text = "0xffffffffc062d000"
.text.unlikely = "0xffffffffc0755300"
__bug_table = "0xffffffffc076c000"
__dyndbg = "0xffffffffc0795728"
__dyndbg_classes = "0xffffffffc07956e8"
__jump_table = "0xffffffffc026a000"
__mcount_loc = "0xffffffffc08f6c90"
__param = "0xffffffffc0924f08"
__patchable_function_entries= "0xffffffffc076e788"
@non_space so nothing changing there, so what about cat /sys/power/mem_sleep again on the working and latest kernels. What snapshot are you at as well?
I am in the newer TW install right now, downloaded on 6/13/25 . . . . Running the newest 6.13.6
# cat /sys/power/mem_sleep
s2idle shallow [deep]
Have to break off on it for now. When I get back to the older TW install Iâll check through the various kernel options and compare them.
On the older install, somewhere over the weekend there was a large 2048 package zypper upgrade, then another 147, and then this morning, another 256?? [from memory] . . . the old one is âcurrent.â ext4.
I did not run any upgrades on the fresh install, because it is a âtempâ OS for testing only.
Back over in the old TW install & booted in the older 6.12.6 kernel, same answer as the newer one:
sudo cat /sys/power/mem_sleep
[sudo] password for root:
s2idle shallow [deep]
Have to reboot to check the 6.13.5 option . . . and then possibly the 6.13.6 in the older TW . . . .
@non_space if the working one is s2idle then for the newer kernels add the kernel boot option mem_sleep_default=s2idle and reboot/test.
1 Like
OK, I just rebooted into 6.13.5 and got the same answer. But, where am I going to add your suggested data? In /etc/default/grub2 ??? Or hit the e key in grubâs advanced options for the 6.13.6 kernel and add that line in??
Which would be a âtempâ edit??
Seemingly you werenât tagged on my previous reply?? Where would I be adding that boot option? In grub menu somewhere? OR in /etc/default/grub2???
In the meanwhile I just rebooted into the newest kernel in the older TW install, the 6.13.6 and ran the requested command and it looks like it is the same as all of the others?
sudo cat /sys/power/mem_sleep
[sudo] password for root:
s2idle shallow [deep]
OK . . . possible âbreakthroughâ in the revival issue . . . the old TW install, booted in 6.13.6 has been suspended and it did âreviveâ the display, 4 times . . . which is one more than one of the kernels reviving 3 times, but not the 4th . . . .
The question is now, would the 5th time be back into the black, or, are we âback in basic functionâ . . .??? It doesnât leap back into the GUI . . . the first time took almost 60 seconds, but each time since it has been faster, so perhaps it is âmachine learningâ at work??
It just goes to show that if you whine long enough the machine gets tired of hearing it, and figures out how to fix itself . . . ?? There were another 78 or so packages to dup into the mix today . . . the âanswerâ might have been included there.
Iâll post back when I have tested it out a few times, if all systems are âgoâ then Iâll mark the thread as âdissolvedâ rather than âsolved.â : - )
OK . . . no . . . the fifth time was not the charm, the display did not wake. So, four is better than three, perhaps the 6.13.7 kernel will bring a more consistent revival . . . ??? At which point 6.14.0 will be back to none???
And so it goes, the vicissitudes and vagaries of life . . . the continuing cycle of âupsâ and âdownsâ . . . relentlessly bringing change, reflected in the joys of open source operating systems . . . .
@non_space Did you try setting it to s2idle for any difference?
I asked a couple of times on the âhow to do that?â and there was no reply.
Where am I going to make that edit? In the grub boot menu or in /etc/default/grub2?
What is the exact command and then I will continue with boot or I have to reboot to get that to show up? I thought that since all of the âcat xxxxâ showed âs2idleâ that that was what we were looking for, already there???
@non_space via YaST Bootloader and add as a kernel option mem_sleep_default=s2idle
Itâs set to deep as the default at present the [deep] you should see [s2idle] instead.
1 Like
OK, Iâll give it a shot in a bit . . . Iâm in Leap 16 at the moment. Iâll post back when I have made the edit and have some evidence to report on. Thanks for the pointers.
OK, I made that edit in YaST Bootloader >Kernel Parameters >Optional Kernel Command Line Parameter . . . changing what was there âquietâ to yr suggestion âmem_sleep_default=s2idleâ . . . and rebooted back into 6.13.6 and did a couple of things in my web browser, then suspended and left the machine.
When I came back the display was black, but the machine was still running, so when I clicked on the keyboard the display responded quickly, but the machine had not stopped running. I repeated the suspend and again the display went black, but the machine continued to run . . . .
The goal is to have the machine and display pause to suspend, AND then with keystroke, both return to function, rather than the machine fires up, but the display does not. I changed the option back to âquiet.â
Other problem, earlier, before making these edits in optional parameter, I suspended 6.13.6 and when I came back and clicked âgoâ the display did not revive, even one time. So the intermittency in function is still tagging along iin the 6.13xxx series.
One other question is that as posted here in the cat mem sleep data, the 612.6 kernel shows the same outcome vis s2idle and it revives and suspends as the designers intended . . . .