IRC channel logs

<sneek>ZhaoM, you have 2 messages!

<sneek>ZhaoM, gnucode says: if you have some free time after adding a birth timestamp to ext2fs...maybe you can add checksumming, snapshots, send/receive, RAID support. You know...the easy filesystem things. :)

<sneek>ZhaoM, youpi says: birth time is not really a priority. See the contributing page, 64bit time would be more useful (we will have to have it by 2038)

<ZhaoM>ok

<ZhaoM>damo22: do you have IDE interface on your laptop which can boot with rumpdisk?

<damo22>no

<damo22>all my disks are sata

<ZhaoM>I suspects there will be some issues when the machine has both SATA controller and IDE interface

<ZhaoM>My T60 has those two

<damo22>i dont think it will be an issue, one will come up as wd0 and another will have a different name

<damo22>both drivers are linked to rumpdisk

<ZhaoM>when I boot qemu with without -M q35 with adding -device ahci,id=ahci, i.e. having SATA controller and IDE interface

<ZhaoM>when poweroff it will generate a lot of device 'timeout reading fsbn' and hang

<ZhaoM>Is there a way to remove the IDE interface and only use SATA controller in qemu?

<ZhaoM>except -M q35

<damo22>i dont know

<damo22>you could hack the default machine configuration

<damo22>-M pc is default

<ZhaoM>Yes I'm looking for how to hack the default configuration

<damo22>edit the source

<damo22>hw/i386/pc.c i think

<ZhaoM>Does it make sense to use host_get_time64? https://git.savannah.gnu.org/cgit/hurd/hurd.git/tree/libdiskfs/node-times.c#n79

<ZhaoM>maptime_read is less accurate than host_get_time64

<youpi>ZhaoM: sure

<ZhaoM>youpi: is this the optimal method? I mean is it *correct*?

<youpi>it's the right way to get a 64b time on gnumach, yes

<ZhaoM>ok

<p4r4D0xum>ZhaoM: https://bpa.st/ENOA

<p4r4D0xum>lspci -nn from my elitebook 2540p

<ZhaoM>sneek later tell youpi I'm thinking about the delay issue. maptime_map() has lower latency than host_get_time64() as the latter is a RPC

<sneek>Okay.

<ZhaoM>Monolithic kernel's system call does not need to consider this issue. The previous hurd developers used vm_map to avoid the delay.

<ZhaoM>But the resolution will be restricted :/.

<ZhaoM>p4r4D0xum: Got it, thanks :)

<p4r4D0xum>ZhaoM: Thank you for takin

<p4r4D0xum>interest

<gnucode>morning friends!

<p4r4D0xum>gnucode: G'mornin'

<youpi>sneek: later tell ZhaoM yes mapping the time is more efficient but not that very convenient, calling the RPC is fine, it's not that expensive

<sneek>Welcome back youpi, you have 1 message!

<sneek>youpi, ZhaoM says: I'm thinking about the delay issue. maptime_map() has lower latency than host_get_time64() as the latter is a RPC

<sneek>Got it.

IRC channel logs

2025-05-01.log