/dev/mem doesn't access all of physical memory on Solaris x86.
At present, the best substitute is the undocumented /dev/xsvc.
Jike Song wrote:
> Hi,
>
> I'm trying to port x86info to Solaris/OpenSolaris. However, I encountered a
> problem wrt. MP tables.
> x86info accepts a "-mp" option, and will dump the MP tables via lseek/read
> /dev/mem.
>
> From the MultiProcessor Specification 1.4, this table should be probed as
> following:
>
> The MP Floating Pointer Structure. This structure contains a physical address
> pointer to the
> MP configuration table and other MP feature information bytes. When present,
> this structure
> indicates that the system conforms to the MP specification. This structure
> must be stored in at
> least one of the following memory locations, because the operating system
> searches for the MP
> floating pointer structure in the order described below:
> a. In the first kilobyte of Extended BIOS Data Area (EBDA), or
> b. Within the last kilobyte of system base memory (e.g., 639K-640K for
> systems with 640
> KB of base memory or 511K-512K for systems with 512 KB of base memory) if
> the
> EBDA segment is undefined, or
> c. In the BIOS ROM address space between 0F0000h and 0FFFFFh.
>
>
> But when you run x86info as root, the reading always fails:
>
> [js226...@git x86info]$ sudo x86info
> x86info v1.21. Dave Jones 2001-2007
> Feedback to <da...@redhat.com>.
>
> readEntry: Bad address
> Found 4 cpus[js226...@git x86info]$
>
>
> I wrote a program to do what the function apic_probe() of x86info project does
>
> /* filename: mem.c
> *
> * This is a silly program does what the function apic_probe() of x86info
> project does:
> *
> * -> open /dev/mem
> * -> seek where the MP config table is
> * -> read 1024 bytes for EBDA
> */
> #include <stdio.h>
> #include <stdint.h>
> #include <stdlib.h>
> #include <sys/types.h>
> #include <sys/stat.h>
> #include <fcntl.h>
> #include <unistd.h>
>
> typedef unsigned long vm_offset_t;
> #define BIOS_SIZE 0x10000
> #define ONE_KBYTE 1024
>
>
>
> int main(void)
> {
> uint16_t segment;
> vm_offset_t target;
> unsigned int buffer[BIOS_SIZE];
>
> int fd = open("/dev/mem", O_RDONLY);
> if (fd == -1) {
> perror("open");
> exit(EXIT_FAILURE);
> }
>
> if (lseek(fd, 0x040e, SEEK_SET) == (off_t)-1 ) {
> perror("lseek");
> exit(EXIT_FAILURE);
> }
>
> if (read(fd, &segment, 2) == -1) {
> perror("read");
> exit(EXIT_FAILURE);
> }
>
> printf("%hx\n", segment);
>
> if (segment) {
> target = (vm_offset_t)segment << 4;
> if (lseek(fd, target, SEEK_SET) == (off_t)-1 ) {
> perror("lseek");
> exit(EXIT_FAILURE);
> }
>
> if (read(fd, buffer, ONE_KBYTE) == -1) {
> perror("read");
> exit(EXIT_FAILURE);
> } else
> printf("read of 1024 bytes of EBDA succeeded\n");
> }
>
>
> return 0;
> }
>
>
> [js226...@git ~]$ sudo ./mem
> 9ec0
> read: Bad address
>
>
> Besides, I have test x86info in both Solaris/Linux on a machine. Linux's
> /dev/mem driver handled this reading correctly.
>
> Any clues?
>
>
>
_______________________________________________
opensolaris-code mailing list
opensolaris-code@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
