On Thu, Sep 22, 2016 at 11:36:40AM +0100, Ian Jackson wrote:
> Ian Jackson writes ("Re: [PATCH v1 3/3] xen-livepatch: If hypervisor is not
> compiled with Livepatching"):
> > Konrad Rzeszutek Wilk writes ("Re: [PATCH v1 3/3] xen-livepatch: If
> > hypervisor is not compiled with Livepatching"):
> > > Why would fprintf mess up 'errno' ?
> > Any libc function is entitled to set errno, if it fails.
> And, I forget the exact rules, but I think most are allowed to trash
> it even if they succeed. I just remember that errno needs to be saved
> across libc calls, if the value is important.
> In practice, for example, the first fprintf on a particular FILE* can
> end up setting errno to ENOTTY as the libc calls isatty() to decide on
> the buffering mode.
Xen-devel mailing list