Re: [PATCH] iwl: fix crash in iwl_dbg_tlv_alloc_trigger

From: Kalle Valo <kvalo@codeaurora.org>
To: Jiri Slaby <jslaby@suse.cz>
Cc: johannes.berg@intel.com, linux-kernel@vger.kernel.org,
	"Dieter Nützel" <Dieter@nuetzel-hh.de>,
	"Emmanuel Grumbach" <emmanuel.grumbach@intel.com>,
	"Luca Coelho" <luciano.coelho@intel.com>,
	"Intel Linux Wireless" <linuxwifi@intel.com>,
	"David S. Miller" <davem@davemloft.net>,
	"Jakub Kicinski" <kuba@kernel.org>,
	linux-wireless@vger.kernel.org, netdev@vger.kernel.org
Subject: Re: [PATCH] iwl: fix crash in iwl_dbg_tlv_alloc_trigger
Date: Fri, 12 Jun 2020 10:55:42 +0300	[thread overview]
Message-ID: <87d064k9a9.fsf@codeaurora.org> (raw)
In-Reply-To: <20200612073800.27742-1-jslaby@suse.cz> (Jiri Slaby's message of "Fri, 12 Jun 2020 09:38:00 +0200")

Jiri Slaby <jslaby@suse.cz> writes:

> The tlv passed to iwl_dbg_tlv_alloc_trigger comes from a loaded firmware
> file. The memory can be marked as read-only as firmware could be
> shared. In anyway, writing to this memory is not expected. So,
> iwl_dbg_tlv_alloc_trigger can crash now:
>
>   BUG: unable to handle page fault for address: ffffae2c01bfa794
>   PF: supervisor write access in kernel mode
>   PF: error_code(0x0003) - permissions violation
>   PGD 107d51067 P4D 107d51067 PUD 107d52067 PMD 659ad2067 PTE 8000000662298161
>   CPU: 2 PID: 161 Comm: kworker/2:1 Not tainted 5.7.0-3.gad96a07-default #1 openSUSE Tumbleweed (unreleased)
>   RIP: 0010:iwl_dbg_tlv_alloc_trigger+0x25/0x60 [iwlwifi]
>   Code: eb f2 0f 1f 00 66 66 66 66 90 83 7e 04 33 48 89 f8 44 8b 46 10 48 89 f7 76 40 41 8d 50 ff 83 fa 19 77 23 8b 56 20 85 d2 75 07 <c7> 46 20 ff ff ff ff 4b 8d 14 40 48 c1 e2 04 48 8d b4 10 00 05 00
>   RSP: 0018:ffffae2c00417ce8 EFLAGS: 00010246
>   RAX: ffff8f0522334018 RBX: ffff8f0522334018 RCX: ffffffffc0fc26c0
>   RDX: 0000000000000000 RSI: ffffae2c01bfa774 RDI: ffffae2c01bfa774
>   RBP: 0000000000000000 R08: 0000000000000004 R09: 0000000000000001
>   R10: 0000000000000034 R11: ffffae2c01bfa77c R12: ffff8f0522334230
>   R13: 0000000001000009 R14: ffff8f0523fdbc00 R15: ffff8f051f395800
>   FS:  0000000000000000(0000) GS:ffff8f0527c80000(0000) knlGS:0000000000000000
>   CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>   CR2: ffffae2c01bfa794 CR3: 0000000389eba000 CR4: 00000000000006e0
>   Call Trace:
>    iwl_dbg_tlv_alloc+0x79/0x120 [iwlwifi]
>    iwl_parse_tlv_firmware.isra.0+0x57d/0x1550 [iwlwifi]
>    iwl_req_fw_callback+0x3f8/0x6a0 [iwlwifi]
>    request_firmware_work_func+0x47/0x90
>    process_one_work+0x1e3/0x3b0
>    worker_thread+0x46/0x340
>    kthread+0x115/0x140
>    ret_from_fork+0x1f/0x40
>
> As can be seen, write bit is not set in the PTE. Read of
> trig->occurrences succeeds in iwl_dbg_tlv_alloc_trigger, but
> trig->occurrences = cpu_to_le32(-1); fails there, obviously.
>
> This is likely because we (at SUSE) use compressed firmware and that is
> marked as RO after decompression (see fw_map_paged_buf).
>
> Fix it by creating a temporary buffer in case we need to change the
> memory.
>
> Signed-off-by: Jiri Slaby <jslaby@suse.cz>
> Reported-by: Dieter Nützel <Dieter@nuetzel-hh.de>
> Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
> Cc: Johannes Berg <johannes.berg@intel.com>
> Cc: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
> Cc: Luca Coelho <luciano.coelho@intel.com>
> Cc: Intel Linux Wireless <linuxwifi@intel.com>
> Cc: Kalle Valo <kvalo@codeaurora.org>
> Cc: "David S. Miller" <davem@davemloft.net>
> Cc: Jakub Kicinski <kuba@kernel.org>
> Cc: linux-wireless@vger.kernel.org
> Cc: netdev@vger.kernel.org
> ---
>  drivers/net/wireless/intel/iwlwifi/iwl-dbg-tlv.c | 16 ++++++++++++++--

The prefix should be "iwlwifi: ", I can fix that.

Luca, should I take this to wireless-drivers?

-- 
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches