> > Hi folks, > > I am looking at a S10 system dump, that has all kernel and user pages > dumped to it. > > I am looking at the stack of a particular kernel thread, that does a > door upcall to a user process acting as the door server. > > Since I know the proc_t address for the user process, I thought of > switching to the user processes's context in MDB, so that I could look > at the userland stack for the process. But on trying that, I get these > errors : > > > 300052003f0::context > debugger context set to proc 300052003f0 > mdb: failed to read /usr/lib/libCstd.so.1:.dynstr: no mapping for address > mdb: failed to read ELF header from /usr/cluster/lib/librgm.so.1: no mapping > for address > mdb: failed to read /lib/libm.so.2:.dynstr: no mapping for address > mdb: failed to read ELF header from /usr/cluster/lib/libscconf.so.1: no > mapping for address > mdb: failed to read ELF header from /usr/cluster/lib/libfe.so.1: no mapping > for address > mdb: failed to read /usr/cluster/lib/libclcomm.so.1:.dynstr: no mapping for > address > mdb: failed to read ELF header from /usr/cluster/lib/libsczones.so.1: no > mapping for address > mdb: failed to read ELF header from /usr/cluster/lib/libscprivip.so.1: no > mapping for address > mdb: failed to read ELF header from /usr/cluster/lib/libdid.so.1: no mapping > for address > mdb: failed to read ELF header from /usr/cluster/lib/libpnm.so.1: no mapping > for address > mdb: failed to read ELF header from /lib/libproc.so.1: no mapping for address > mdb: failed to read ELF header from /lib/librtld_db.so.1: no mapping for > address > mdb: failed to read ELF header from /lib/libelf.so.1: no mapping for address > mdb: failed to read ELF header from /lib/libctf.so.1: no mapping for address > mdb: failed to read ELF header from /lib/nss_files.so.1: no mapping for > address > > > > ::status > debugging PID 981 (32-bit) in kernel crash dump > executable file: /usr/cluster/lib/sc/rgmd > > > > > I am not sure if these errors signify something wrong with the dump. > Do these sort of errors seem familiar to anyone? > > Thanks & Regards, > Sambit
Most likely the contents of those pages are in swap space and thus not part of the dump. You should be able to confirm this by looking at other portions of the process's address space and confirming you can read at least something (e.g. using ::dis or ::dump or something). -Mike -- Mike Shapiro, Solaris Kernel Development. blogs.sun.com/mws/