From mboxrd@z Thu Jan 1 00:00:00 1970 From: Al Viro Subject: Re: [PATCH 06/24] fs: add a kern_chdir helper Date: Mon, 20 Jul 2020 19:17:06 +0100 Message-ID: <20200720181706.GR2786714@ZenIV.linux.org.uk> References: <20200720155902.181712-1-hch@lst.de> <20200720155902.181712-7-hch@lst.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20200720155902.181712-7-hch@lst.de> Sender: linux-kernel-owner@vger.kernel.org To: Christoph Hellwig Cc: Linus Torvalds , Greg Kroah-Hartman , "Rafael J. Wysocki" , linux-kernel@vger.kernel.org, linux-raid@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-api@vger.kernel.org List-Id: linux-raid.ids On Mon, Jul 20, 2020 at 05:58:44PM +0200, Christoph Hellwig wrote: > Add a simple helper for a chdir with a kernelspace name and use it in the > early init code instead of relying on the implicit set_fs(KERNEL_DS) > there. Remove the now unused ksys_chdir. FWIW, my problem with that is that you are making it first-class primitive. And I don't see any valid use for it - existing one is really "we have a userland process that would be making syscalls if we only could have it done right; the only reason we can't is that the infrastructure to build the binary would be hell to maintain in the tree". It's *NOT* "kernel code that has a valid reason to get the syscall functionality".