From: Emmanuel Fleury <[EMAIL PROTECTED]> > What is suprising me is that the kernel should take care of the > liveness of the system and it seems to not detect this problem (I > mean the distribution of cpu time should be fair among processes). > > I can add that I noticed that even if there is no more ssh, no more > inetd, and so on, the ping is responding. Which means that the very low > level kernel features are still active. > > So, this problem *could* be also related to a Linux kernel problem in > > the SMP part... (I mean, you can blow the system like this only with > sound driver, do you?).
Like any driver, it owns the cpu until finished or until driver code gives cpu away. The driver can only be pre-empted by interrupts and such. The kernel is not monitoring itself, it's monitoring user space processes... The alsamixer is probably making a sys-call (after that we are running in kernel space) and that does not return - and - seems to hold the cpu... As pings are working, interrupts and such are working. - They pre-empt the spinning driver but return cpu back to it when finished... http://kernelnewbies.org/documents/kdoc/kernel-hacking/basic-players.html Regards, Robert _______________________________________________ Alsa-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/alsa-devel
