From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jarkko Sakkinen Subject: Re: [tpmdd-devel] [PATCH v2 2/7] tpm: two-phase chip management functions Date: Thu, 9 Oct 2014 12:07:47 +0300 Message-ID: <20141009090747.GA24880@intel.com> References: <1412701277-27794-1-git-send-email-jarkko.sakkinen@linux.intel.com> <1412701277-27794-3-git-send-email-jarkko.sakkinen@linux.intel.com> <20141007175017.GA10432@obsidianresearch.com> <20141007222814.GA7262@intel.com> <20141007223442.GA3014@obsidianresearch.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20141007223442.GA3014-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org> Sender: linux-api-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Jason Gunthorpe Cc: Peter Huewe , Ashley Lai , Marcel Selhorst , linux-api-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, tpmdd-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: linux-api@vger.kernel.org On Tue, Oct 07, 2014 at 04:34:42PM -0600, Jason Gunthorpe wrote: > On Wed, Oct 08, 2014 at 01:28:14AM +0300, Jarkko Sakkinen wrote: > > > > > @@ -714,15 +709,10 @@ static int tpm_tis_i2c_remove(struct i2c_client *client) > > > > struct tpm_chip *chip = tpm_dev.chip; > > > > release_locality(chip, chip->vendor.locality, 1); > > > > > > > > - /* close file handles */ > > > > - tpm_dev_vendor_release(chip); > > > > - > > > > /* remove hardware */ > > > > tpm_remove_hardware(chip->dev); > > > > > > Wrong ordering here, tpm_remove_hardware should always be first - > > > drivers should not tear down internal state before calling it, so > > > release_locality should be second. > > > > > > Noting that since we use devm the kfree will not happen until > > > remove returns, so the chip pointer is still valid. > > > > Should I fix this ordering? I was thinking to focus putting proper > > patterns in place only in tpm_tis and tpm_crb because they are the > > that I'm able to test easily and then they can work as guideline for > > other drivers. > > I think since this patch is already touching this function there is > no reason not to make it be correct (especially since it was noticed) > > The rest can wait till we globally replace tpm_remove_hardware with > tpm_unregister - at that time the ordering can be audited and > checked. > > Then the drivers will be clean and the core can finally be fixed. This makes sense. I'll also document this. And I decided to completely wipe old tpm_register/remove_hardware() completely from v3 because they only cause confusion. I pushed patch that should implement fix for the ordering into tpm2-v2 branch: https://github.com/jsakkine/linux-tpm2/commit/63ab650fa6f8dddd95100869e50275801d7d9360 > Jason /Jarkko