NAME

errfile − error-log file format

DESCRIPTION

When hardware errors are detected by the system, an error record is generated and passed to the error-logging daemon for recording in the error log for later analysis.  The default error log is /usr/adm/errfile. 

The format of an error record depends on the type of error that was encountered.  Every record, however, has a header with the following format:

struct errhdr {
int e_type; /∗ record type ∗/
int e_len; /∗ bytes in record (inc hdr) ∗/
time_t e_time; /∗ time of day ∗/
};

The permissible record types are as follows:

#define E_GOTS 010 /∗ Start for UNIX 3.0∗/
#define E_GORT 011 /∗ Start for UNIX/RT ∗/
#define E_STOP 012 /∗ Stop ∗/
#define E_TCHG 013 /∗ Time change ∗/
#define E_CCHG 014 /∗ Configuration change ∗/
#define E_BLK 020 /∗ Block device error ∗/
#define E_STRAY 030 /∗ Stray interrupt ∗/
#define E_PRTY 031 /∗ Memory parity ∗/

Some records in the error file are of an administrative nature.  These include the startup record that is entered into the file when logging is activated, the stop record that is written if the daemon is terminated “gracefully”, and the time-change record that is used to account for changes in the system’s time-of-day.  These records have the following formats:

struct estart {
struct errhdr e_hdr; /∗ record header ∗/
int e_cpu; /∗ CPU type ∗/
int e_mmr3; /∗ contents mem mgmt reg 3 ∗/
long e_syssize; /∗ 11/70 system memory size ∗/
int e_bconf; /∗ block dev configuration ∗/
};

struct eend {
struct errhdr e_hdr; /∗ record header ∗/
};

struct etimchg {
struct errhdr e_hdr; /∗ record header ∗/
time_t e_ntime; /∗ new time ∗/
};

Stray interrupts cause a record with the following format to be logged in the file:

struct estray {
struct errhdr e_hdr; /∗ record header ∗/
physadr e_saddr; /∗ stray loc or device addr ∗/
int e_sbacty; /∗ active block devices ∗/
};

Memory subsystem error on 11/70 processors cause the following record to be generated:

struct eparity {
struct errhdr e_hdr; /∗ record header ∗/
int e_parreg[4]; /∗ memory subsys registers ∗/
};

Error records for block devices have the following format:

struct eblock {
struct errhdr e_hdr; /∗ record header ∗/
dev_t e_dev; /∗ "true" major + minor dev no ∗/
physadr e_regloc; /∗ controller address ∗/
int e_bacty; /∗ other block I/O activity ∗/
struct iostat {
   long io_ops; /∗ number read/writes ∗/
   long io_misc; /∗ number "other" operations ∗/
   unsigned io_unlog; /∗ number unlogged errors ∗/
} e_stats;
int e_bflags; /∗ read/write, error, etc ∗/
int e_cyloff; /∗ logical dev start cyl ∗/
daddr_t e_bnum; /∗ logical block number ∗/
unsigned e_bytes; /∗ number bytes to transfer ∗/
long e_memadd; /∗ buffer memory address ∗/
unsigned e_rtry; /∗ number retries ∗/
int e_nreg; /∗ number device registers ∗/
};

The following values are used in the e_bflags word:

#define E_WRITE 0 /∗ write operation ∗/
#define E_READ 1 /∗ read operation ∗/
#define E_NOIO 02 /∗ no I/O pending ∗/
#define E_PHYS 04 /∗ physical I/O ∗/
#define E_MAP 010 /∗ Unibus map in use ∗/
#define E_ERROR 020 /∗ I/O failed ∗/

The “true” major device numbers that identify the failing device are as follows:

#define RK0 0
#define RP0 1
#define RF0 2
#define TM0 3
#define TC0 4
#define HP0 5
#define HT0 6
#define HS0 7

SEE ALSO

errdemon(1M). 

May 16, 1980