vmalloc and pkmap overlap on 2.4

The vmalloc pool runs from 0xf800 to FIXADDR_START, which is a few pages below the top of virtual address space. The pkmap area is from 0xfe00 to 0xfe40, which fits within the vmalloc pool. There does not seem to be any mechanism to prevent these two from stepping on each other's toe

System.map with symbols from discarded sections

Quite a few functions in the 2.3 kernels and up are marked as __exit. This puts the functions in the .text.exit section that is marked as DISCARD in vmlinux.lds. It turns out that if the function is static, ld never puts it into the symbol table of vmlinux; however, if the function is global, ld

Fixing random corruption in raw IO on 2.2.x kernel with bigmem enabled

efore a bigmem page in the array, which in turn leads to the writing of uninitialized (garbage) pages to the raw device. The corruption happens rather infrequently. It has been reproduced both with the SAP DB and without. It appears that intense paging activity increases its chance of occurr