On Tue, 2006-01-17 at 12:56 +0100, Hendrik Post wrote: > Hi! > > I found several places in the kernel source code where kmalloc is called > without any check if the return value is null. Is this a serious matter, > e.g. should such code be fixed? sometimes > > My intuition is that otherwise this could lead to code defreferencing a > null pointer, though it is not likely that kmalloc fails... > > Is it reasonable to enforce such checks for errors for all modules? for modules, yes. There are a set of known cases which ONLY happen during really early boot, and for which no reasonable recovery is possible. (eg if you run out of memory before even having started userspace, your system isn't going to be any good anyway, and during that phase of booting you NEED to have certain things. Like a root filesystem, or like things internal to the VM). Those are known cases and ok as is. But in "drivers" which can be a module, it generally is a bug that wants fixing yes. -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
