I had a look at the crash dump file. Looking into the invocation of the pcieadm command whose thread initiated the panic in the dump > ::ps ! grep pciead R 7390 7353 7390 7353 0 0x4a004000 fffffe58cc671218 pcieadm > fffffe58cc671218::print proc_t ! grep psar u_psargs = [ "/usr/lib/pci/pcieadm show-cfgspace -d 57/0/0" ] > > > fffffe58cc671218::pfiles FD TYPE VNODE INFO 0 CHR fffffe58fc5aca80 /dev/pts/2 1 CHR fffffe58fc5aca80 /dev/pts/2 2 CHR fffffe58fc5aca80 /dev/pts/2 3 CHR fffffe58e443f340 /devices/pci@0,0:reg > Neither the arguments to the "pcieadm" command nor the fd's associated with the pcieadm process has any reference to /tmp . As you mentioned in one of the previous emails (about encountering the panic w/o /tmp) , this panic did not involve /tmp. -- > Your first email on this thread mentions a panic due to PCI error > panic message: pcieb-2: PCI(-X) Express Fatal Error. (0x43) > > Is that resolved? I think? I uttered `fmadm repair` to it and FMA stopped complaining. But that's all, I can reproduce it w/o involving /tmp... --- I think the panics triggered by pcieb_intr_handler are not non-pcie related. Pramod On Wed, Mar 13, 2024 at 10:55 PM Dan McDonald wrote: > Sorry for the latency on this. Exchange or Mail.app put your last email > in junk. > > https://kebe.com/~danmcd/webrevs/pcie-vmdump.3 < > https://kebe.com/~danmcd/webrevs/pcie-vmdump.3> > > MD5 == 1036261a0c22e5f213e07e4b317b9001 > > Dan > > > ------------------------------------------ > illumos: illumos-developer > Permalink: > https://illumos.topicbox.com/groups/developer/T32516b4f237e0abc-Mc84ee0001adef12d744a0dce > Delivery options: > https://illumos.topicbox.com/groups/developer/subscription >