Linux-Sgx Archive on lore.kernel.org
 help / color / Atom feed
From: Sean Christopherson <sean.j.christopherson@intel.com>
To: Borislav Petkov <bp@alien8.de>
Cc: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>,
	linux-kernel@vger.kernel.org, x86@kernel.org,
	linux-sgx@vger.kernel.org, akpm@linux-foundation.org,
	dave.hansen@intel.com, nhorman@redhat.com, npmccallum@redhat.com,
	serge.ayoun@intel.com, shay.katz-zamir@intel.com,
	haitao.huang@intel.com, andriy.shevchenko@linux.intel.com,
	tglx@linutronix.de, kai.svahn@intel.com, josh@joshtriplett.org,
	luto@kernel.org, kai.huang@intel.com, rientjes@google.com,
	cedric.xing@intel.com
Subject: Re: [PATCH v22 09/24] x86/sgx: Add functions to allocate and free EPC pages
Date: Mon, 7 Oct 2019 07:50:11 -0700
Message-ID: <20191007145011.GA18016@linux.intel.com> (raw)
In-Reply-To: <20191005164408.GB25699@zn.tnic>

On Sat, Oct 05, 2019 at 06:44:08PM +0200, Borislav Petkov wrote:
> On Tue, Sep 03, 2019 at 05:26:40PM +0300, Jarkko Sakkinen wrote:
> > +/**
> > + * __sgx_free_page - Free an EPC page
> > + * @page:	pointer a previously allocated EPC page
> > + *
> > + * EREMOVE an EPC page and insert it back to the list of free pages.
> > + *
> > + * Return:
> > + *   0 on success
> > + *   SGX error code if EREMOVE fails
> > + */
> > +int __sgx_free_page(struct sgx_epc_page *page)
> > +{
> > +	struct sgx_epc_section *section = sgx_epc_section(page);
> > +	int ret;
> 
> Shouldn't you be grabbing the lock here already?
> 
> Or nothing can happen to that page from another thread after you
> ENCLS[EREMOVE] it and before it is added to the ->page_list of the
> section?

The caller is responsible for ensuring EREMOVE can be safely executed,
e.g. by holding the enclave's lock.

For many ENCLS leafs, EREMOVE included, the CPU requires exclusive access
to the SGX Enclave Control Structures (SECS)[*] and will signal a #GP if
a different logical CPU is already executing an ENCLS leaf that requires
exclusive SECS access.  The SGX subsystem uses a per-enclave mutex to
serialize such ENCLS leafs, among other things.

[*] The SECS is a per-enclave page that resides in the EPC and can only be
    directly accessed by the CPU.  It's used to track metadata about the
    enclave, e.g. number of child pages, base, size, etc...

> > +
> > +	ret = __eremove(sgx_epc_addr(page));
> > +	if (ret)
> > +		return ret;
> > +
> > +	spin_lock(&section->lock);
> > +	list_add_tail(&page->list, &section->page_list);
> > +	section->free_cnt++;
> > +	spin_unlock(&section->lock);
> > +
> > +	return 0;
> > +}
> > +EXPORT_SYMBOL_GPL(__sgx_free_page);
> > +
> > +/**
> > + * sgx_free_page - Free an EPC page and WARN on failure
> > + * @page:	pointer to a previously allocated EPC page
> > + *
> > + * EREMOVE an EPC page and insert it back to the list of free pages, and WARN
> > + * if EREMOVE fails.  For use when the call site cannot (or chooses not to)
> > + * handle failure, i.e. the page is leaked on failure.
> > + */
> > +void sgx_free_page(struct sgx_epc_page *page)
> > +{
> > +	int ret;
> > +
> > +	ret = __sgx_free_page(page);
> > +	WARN(ret > 0, "sgx: EREMOVE returned %d (0x%x)", ret, ret);
> 
> That will potentially flood dmesg. Why are we even accommodating such
> callers? They either handle the error or they don't get to alloc EPC
> pages. There's also __must_check with which you can enforce the error
> code checking or we simply don't allow not handling failure. Fullstop.

It was deliberately left as a WARN as having multiple stack traces has
been very helpful for triaging and debugging software bugs.  I agree that
it can and should be changed to a WARN_ONCE() for upstream.

As for why it exists at all, the WARN will only fire if there is a kernel
and/or CPU bug.  In the vast majority of cases, EREMOVE is guaranteed to
be successful (assuming no bugs).  In those situations, if EREMOVE fails
there really isn't a sane option other to WARN and leak the page, e.g.
the kernel can't override the CPUs protection to forcefully EREMOVE the
page.

The non-warn variant, __sgx_free_page(), is provided for the (currently)
one case where EREMOVE failure is expected/tolerable (opportunstically
freeing EPC pages when the enclave is killed).

  reply index

Thread overview: 102+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-09-03 14:26 [PATCH v22 00/24] Intel SGX foundations Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 01/24] x86/cpufeatures: x86/msr: Add Intel SGX hardware bits Jarkko Sakkinen
2019-09-24 15:28   ` Borislav Petkov
2019-09-24 16:11     ` Sean Christopherson
2019-09-24 16:25       ` Borislav Petkov
2019-09-03 14:26 ` [PATCH v22 02/24] x86/cpufeatures: x86/msr: Intel SGX Launch Control " Jarkko Sakkinen
2019-09-24 15:52   ` Borislav Petkov
2019-09-24 20:22     ` Sean Christopherson
2019-09-25  8:51       ` Borislav Petkov
2019-09-25 17:18         ` Sean Christopherson
2019-09-25 18:31           ` Borislav Petkov
2019-09-25 19:08             ` Sean Christopherson
2019-09-27 16:11           ` Jarkko Sakkinen
2019-09-25 14:09     ` Jarkko Sakkinen
2019-09-25 14:10       ` Jarkko Sakkinen
2019-09-25 14:38         ` Jarkko Sakkinen
2019-09-25 15:19       ` Borislav Petkov
2019-09-25 16:49         ` Sean Christopherson
2019-09-25 17:28           ` Borislav Petkov
2019-09-25 18:18             ` Sean Christopherson
2019-09-03 14:26 ` [PATCH v22 03/24] x86/mm: x86/sgx: Signal SIGSEGV with PF_SGX Jarkko Sakkinen
2019-09-24 16:04   ` Borislav Petkov
2019-09-25 14:16     ` Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 04/24] x86/cpu/intel: Detect SGX supprt Jarkko Sakkinen
2019-09-24 16:13   ` Borislav Petkov
2019-09-24 17:43     ` Sean Christopherson
2019-09-24 18:21       ` Borislav Petkov
2019-09-25 14:46         ` Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 05/24] x86/sgx: Add ENCLS architectural error codes Jarkko Sakkinen
2019-09-27 10:20   ` Borislav Petkov
2019-09-27 16:08     ` Jarkko Sakkinen
2019-09-27 17:20       ` Sean Christopherson
2019-10-01 20:23         ` Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 06/24] x86/sgx: Add SGX microarchitectural data structures Jarkko Sakkinen
2019-09-27 16:27   ` Borislav Petkov
2019-10-01 19:10     ` Jarkko Sakkinen
2019-10-01 20:39     ` Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 07/24] x86/sgx: Add wrappers for ENCLS leaf functions Jarkko Sakkinen
2019-10-04  9:45   ` Borislav Petkov
2019-10-04 18:56     ` Jarkko Sakkinen
2019-10-08  4:04     ` Sean Christopherson
2019-10-08  7:18       ` Borislav Petkov
2019-10-08 13:35         ` Sean Christopherson
2019-10-08 14:56           ` Borislav Petkov
2019-09-03 14:26 ` [PATCH v22 08/24] x86/sgx: Enumerate and track EPC sections Jarkko Sakkinen
2019-10-05  9:26   ` Borislav Petkov
2019-10-07 11:58     ` Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 09/24] x86/sgx: Add functions to allocate and free EPC pages Jarkko Sakkinen
2019-10-05 16:44   ` Borislav Petkov
2019-10-07 14:50     ` Sean Christopherson [this message]
2019-10-08  9:09       ` Borislav Petkov
2019-10-08 13:31         ` Sean Christopherson
2019-10-07 17:55     ` Jarkko Sakkinen
2019-10-07 18:09       ` Borislav Petkov
2019-09-03 14:26 ` [PATCH v22 10/24] x86/sgx: Add sgx_einit() for wrapping ENCLS[EINIT] Jarkko Sakkinen
2019-10-08 17:30   ` Borislav Petkov
2019-10-08 17:45     ` Sean Christopherson
2019-10-08 17:46       ` Sean Christopherson
2019-10-08 17:53         ` Borislav Petkov
2019-09-03 14:26 ` [PATCH v22 11/24] mm: Introduce vm_ops->may_mprotect() Jarkko Sakkinen
2019-10-08 17:41   ` Borislav Petkov
2019-09-03 14:26 ` [PATCH v22 12/24] x86/sgx: Linux Enclave Driver Jarkko Sakkinen
2019-10-08 17:59   ` Borislav Petkov
2019-10-08 18:17     ` Sean Christopherson
2019-10-08 19:19       ` Borislav Petkov
2019-09-03 14:26 ` [PATCH v22 13/24] x86/sgx: Add provisioning Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 14/24] x86/sgx: Add a page reclaimer Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 15/24] x86/sgx: ptrace() support for the SGX driver Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 16/24] x86/vdso: Add support for exception fixup in vDSO functions Jarkko Sakkinen
2019-10-02 23:18   ` Jarkko Sakkinen
2019-10-02 23:45     ` Jarkko Sakkinen
2019-10-04  0:03     ` Sean Christopherson
2019-10-04 18:49       ` Jarkko Sakkinen
2019-10-04  0:15     ` Sean Christopherson
2019-10-04 18:52       ` Jarkko Sakkinen
2019-10-05 15:54         ` Sean Christopherson
2019-10-07  7:57           ` Jarkko Sakkinen
2019-10-07  8:10             ` Jarkko Sakkinen
2019-10-07 12:04               ` Jarkko Sakkinen
2019-10-08  4:54                 ` Sean Christopherson
2019-10-05 18:39         ` Sean Christopherson
2019-10-07  8:01           ` Jarkko Sakkinen
2019-10-06 23:38         ` Jarkko Sakkinen
2019-10-06 23:40           ` Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 17/24] x86/fault: Add helper function to sanitize error code Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 18/24] x86/traps: Attempt to fixup exceptions in vDSO before signaling Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 19/24] x86/vdso: Add __vdso_sgx_enter_enclave() to wrap SGX enclave transitions Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 20/24] selftests/x86: Add a selftest for SGX Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 21/24] selftests/x86: Recurse into subdirectories Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 22/24] x86/sgx: Update MAINTAINERS Jarkko Sakkinen
2019-09-03 14:26 ` [PATCH v22 23/24] docs: x86/sgx: Document microarchitecture Jarkko Sakkinen
2019-09-27 18:15   ` Randy Dunlap
2019-09-03 14:26 ` [PATCH v22 24/24] docs: x86/sgx: Document kernel internals Jarkko Sakkinen
2019-09-27 17:07   ` Randy Dunlap
2019-10-01 19:34     ` Jarkko Sakkinen
2019-09-13 20:38 ` [PATCH v22 00/24] Intel SGX foundations Dave Hansen
2019-09-14 13:41   ` Jarkko Sakkinen
2019-09-14 15:32     ` Dave Hansen
2019-09-16  5:23       ` Jarkko Sakkinen
2019-09-24 17:20         ` Andy Lutomirski
2019-09-25 14:32           ` Jarkko Sakkinen
2019-10-02 23:42             ` Jarkko Sakkinen

Reply instructions:

You may reply publically to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20191007145011.GA18016@linux.intel.com \
    --to=sean.j.christopherson@intel.com \
    --cc=akpm@linux-foundation.org \
    --cc=andriy.shevchenko@linux.intel.com \
    --cc=bp@alien8.de \
    --cc=cedric.xing@intel.com \
    --cc=dave.hansen@intel.com \
    --cc=haitao.huang@intel.com \
    --cc=jarkko.sakkinen@linux.intel.com \
    --cc=josh@joshtriplett.org \
    --cc=kai.huang@intel.com \
    --cc=kai.svahn@intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-sgx@vger.kernel.org \
    --cc=luto@kernel.org \
    --cc=nhorman@redhat.com \
    --cc=npmccallum@redhat.com \
    --cc=rientjes@google.com \
    --cc=serge.ayoun@intel.com \
    --cc=shay.katz-zamir@intel.com \
    --cc=tglx@linutronix.de \
    --cc=x86@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-Sgx Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-sgx/0 linux-sgx/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-sgx linux-sgx/ https://lore.kernel.org/linux-sgx \
		linux-sgx@vger.kernel.org
	public-inbox-index linux-sgx

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-sgx


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git