netdev - Re: [PATCH net-next v2 6/9] ice: use <linux/packing.h> for Tx and Rx queue context data

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <62387bab-f42a-4981-9664-76c439e2aadb@intel.com>
Date: Tue, 29 Oct 2024 16:32:34 -0700
From: Jacob Keller <jacob.e.keller@...el.com>
To: Daniel Machon <daniel.machon@...rochip.com>, Dan Carpenter
	<dan.carpenter@...aro.org>
CC: Vladimir Oltean <olteanv@...il.com>, Andrew Morton
	<akpm@...ux-foundation.org>, Eric Dumazet <edumazet@...gle.com>, "Jakub
 Kicinski" <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>, Tony Nguyen
	<anthony.l.nguyen@...el.com>, Przemek Kitszel <przemyslaw.kitszel@...el.com>,
	Masahiro Yamada <masahiroy@...nel.org>, <linux-kernel@...r.kernel.org>,
	<netdev@...r.kernel.org>
Subject: Re: [PATCH net-next v2 6/9] ice: use <linux/packing.h> for Tx and Rx
 queue context data



On 10/29/2024 3:09 PM, Jacob Keller wrote:
> 
> 
> On 10/29/2024 7:50 AM, Daniel Machon wrote:
>> Hi Jacob,
> 
>>> +/**
>>> + * ice_pack_rxq_ctx - Pack Rx queue context into a HW buffer
>>> + * @ctx: the Rx queue context to pack
>>> + * @buf: the HW buffer to pack into
>>> + *
>>> + * Pack the Rx queue context from the CPU-friendly unpacked buffer into its
>>> + * bit-packed HW layout.
>>> + */
>>> +static void ice_pack_rxq_ctx(const struct ice_rlan_ctx *ctx,
>>> +                            ice_rxq_ctx_buf_t *buf)
>>> +{
>>> +       CHECK_PACKED_FIELDS_20(ice_rlan_ctx_fields, ICE_RXQ_CTX_SZ);
>>> +       BUILD_BUG_ON(sizeof(*buf) != ICE_RXQ_CTX_SZ);
>>> +
>>> +       pack_fields(buf, sizeof(*buf), ctx, ice_rlan_ctx_fields,
>>> +                   QUIRK_LITTLE_ENDIAN | QUIRK_LSW32_IS_FIRST);
>>> +}
>>> +
>>
>> FWIW, I noticed that smatch bails out checking all the CHECK_PACKED_FIELDS_*
>> variants >= 20, with the warning:
>>
>> ice_common.c:1486 ice_pack_txq_ctx() parse error: OOM: 3000148Kb sm_state_count = 413556
>> ice_common.c:1486 ice_pack_txq_ctx() warn: Function too hairy.  No more merges.
>> ice_common.c:1486 ice_pack_txq_ctx() parse error: Function too hairy.  Giving up. 43 second
>>
> 
> We might need to wrap these checks to become no-ops when running under
> such a checker. It looks like the parser doesn't like the size of the
> macros?
> 
>> Maybe this can just be ignored .. not sure :-)
>>
> 
> I would prefer if we found a way to at least silence it, rather than
> straight up ignore it.
> 
> I am not that familiar with smatch. Let me see if there's an obvious way
> we can handle this.
> 
>From the look of this, smatch is running out of memory trying to keep
track of a number of possible states. Likely the depth of the macro is
too high once we get beyond 20.

I think the simplest solution would be to disable these macros under
__CHECKER__, and possibly even write our own checker in smatch which can
cover this case..

@Dan Carpenter:

Any chance you could provide guidance here?

Thanks,
Jake