I'm currently sitting in KMDB looking at a pattern of data corruption. ::kgrep with a mask can identify all of the kernel locations which share this corruption pattern (unfortunately it doesn't show up elsehwere). Now I'd like to search all of physical memory, but I'm having issues.
[16]> 1f2f46990,3\K 0x1f2f46990: c801300000400000 cc01300000400000 d001300000400000 So I want to search for 300000400000 and I'd expect to get 3 hits at these addresses, however that doesn't appear to be the case. [16]> 1f2f46990,3\M 300000400000 ffffffffffff 0x1f2f46990 So it give me a pointer to the first location, but not the second and third. Is it possible that the search will just give me the first occurence and then quit? And then I'd have to increment by 8 and start again? If so, is there any sort of pipe sequence which can automate this? In addition, I don't understand how the mask is supposed to work. If I do: [16]> 1f2f46990,3\M c801300000400000 ffffffffffff It just hangs. This seems like odd behavior. At least ^C gets me out of this. First is there a way to do what I want to do? If so, what am I doing wrong? I really don't want to dump out all of physical memory and then postprocess the information since that will take quite a bit of time on an 8G system :( Thanks This message posted from opensolaris.org
