Hello, Sorry for late reply.
y...@mwd.biglobe.ne.jp (YAMAMOTO Takashi) wrote: > >>> > >>> how about the attached patch? > >> > >> Consider offset = (INT64_MAX - PAGE_SIZE) and len = PAGE_SIZE. That > >> would still panic.. > > > > heh, right. > > > > then, how about: > > if (round_page(offset) < trunc_page(endoffset)) { > > && offset < round_page(offset) I think that should be correct, except there is off-by-one (since offset at PAGE_SIZE boundary is valid). Should be: ... && offset <= round_page(offset) Do you want to commit this? -- Mindaugas