Hi Jérôme,
Answers inline.
[Add Krzysztof in Cc:]
W dniu 15.09.2022 o 15:39, Jérôme Pouiller pisze:
> [Add Peter in Cc:]
>
> On Thursday 15 September 2022 15:14:45 CEST Lech Perczak wrote:
> > From: Paweł Lenkow <pawel.lenkow@xxxxxxxxxxxxxxx>
> >
> > During our testing of WFM200 module over SDIO on i.MX6Q-based platform,
> > we discovered a memory corruption on the system, tracing back to the wfx
> > driver. Using kfence, it was possible to trace it back to the root
> > cause, which is hw->max_rates set to 8 in wfx_init_common,
> > while the maximum defined by IEEE80211_TX_TABLE_SIZE is 4.
> >
> > This causes array out-of-bounds writes during updates of the rate table,
> > as seen below:
> >
> > BUG: KFENCE: memory corruption in kfree_rcu_work+0x320/0x36c
> >
> > Corrupted memory at 0xe0a4ffe0 [ 0x03 0x03 0x03 0x03 0x01 0x00 0x00
> > 0x02 0x02 0x02 0x09 0x00 0x21 0xbb 0xbb 0xbb ] (in kfence-#81):
> > kfree_rcu_work+0x320/0x36c
> > process_one_work+0x3ec/0x920
> > worker_thread+0x60/0x7a4
> > kthread+0x174/0x1b4
> > ret_from_fork+0x14/0x2c
> > 0x0
> >
> > kfence-#81: 0xe0a4ffc0-0xe0a4ffdf, size=32, cache=kmalloc-64
> >
> > allocated by task 297 on cpu 0 at 631.039555s:
> > minstrel_ht_update_rates+0x38/0x2b0 [mac80211]
> > rate_control_tx_status+0xb4/0x148 [mac80211]
> > ieee80211_tx_status_ext+0x364/0x1030 [mac80211]
> > ieee80211_tx_status+0xe0/0x118 [mac80211]
> > ieee80211_tasklet_handler+0xb0/0xe0 [mac80211]
> > tasklet_action_common.constprop.0+0x11c/0x148
> > __do_softirq+0x1a4/0x61c
> > irq_exit+0xcc/0x104
> > call_with_stack+0x18/0x20
> > __irq_svc+0x80/0xb0
> > wq_worker_sleeping+0x10/0x100
> > wq_worker_sleeping+0x10/0x100
> > schedule+0x50/0xe0
> > schedule_timeout+0x2e0/0x474
> > wait_for_completion+0xdc/0x1ec
> > mmc_wait_for_req_done+0xc4/0xf8
> > mmc_io_rw_extended+0x3b4/0x4ec
> > sdio_io_rw_ext_helper+0x290/0x384
> > sdio_memcpy_toio+0x30/0x38
> > wfx_sdio_copy_to_io+0x88/0x108 [wfx]
> > wfx_data_write+0x88/0x1f0 [wfx]
> > bh_work+0x1c8/0xcc0 [wfx]
> > process_one_work+0x3ec/0x920
> > worker_thread+0x60/0x7a4
> > kthread+0x174/0x1b4
> > ret_from_fork+0x14/0x2c 0x0
> >
> > Limit hw->max_rates to not exceed IEEE80211_TX_RATE_TABLE_SIZE (4).
> >
> > To bring back previous value, the global table size limit needs to be
> > increased beforehand in mac80211.h, or by limiting the iteration count
> > in minstrel_ht_update_rates against IEEE80211_TX_RATE_TABLE_SIZE as
> > well.
> >
> > Fixes: e16e7f0716a6 ("staging: wfx: instantiate mac80211 data")
>
> I think the issue has been introduced by ee0e16ab756a ("mac80211:
> minstrel_ht: fill all requested rates").
>
>
> > Cc: Jérôme Pouiller <jerome.pouiller@xxxxxxxxxx>
> > Cc: Kalle Valo <kvalo@xxxxxxxxxx>
> > Link: https://urldefense.com/v3/__https://lore.kernel.org/all/12e5adcd-8aed-f0f7-70cc-4fb7b656b829@xxxxxxxxxxxxxxx/__;!!N30Cs7Jr!ReVaYMRjWoJzG95KRgrZTGAw0bmt5lnLLpRkt574SRvIoKLD2xl53YKUiLpN4PfXpjSLIQ9KvgVy9Wi4jeJE8axP9M4Odgk$ <https://urldefense.com/v3/__https://lore.kernel.org/all/12e5adcd-8aed-f0f7-70cc-4fb7b656b829@xxxxxxxxxxxxxxx/__;!!N30Cs7Jr!ReVaYMRjWoJzG95KRgrZTGAw0bmt5lnLLpRkt574SRvIoKLD2xl53YKUiLpN4PfXpjSLIQ9KvgVy9Wi4jeJE8axP9M4Odgk$>
> >
> > Signed-off-by: Paweł Lenkow <pawel.lenkow@xxxxxxxxxxxxxxx>
> > Signed-off-by: Lech Perczak <lech.perczak@xxxxxxxxxxxxxxx>
> > ---
> > drivers/net/wireless/silabs/wfx/main.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/net/wireless/silabs/wfx/main.c b/drivers/net/wireless/silabs/wfx/main.c
> > index 84d82ddded56..7463fe4b5cae 100644
> > --- a/drivers/net/wireless/silabs/wfx/main.c
> > +++ b/drivers/net/wireless/silabs/wfx/main.c
> > @@ -273,7 +273,7 @@ struct wfx_dev *wfx_init_common(struct device *dev, const struct wfx_platform_da
> > hw->vif_data_size = sizeof(struct wfx_vif);
> > hw->sta_data_size = sizeof(struct wfx_sta_priv);
> > hw->queues = 4;
> > - hw->max_rates = 8;
> > + hw->max_rates = 4;
> > hw->max_rate_tries = 8;
> > hw->extra_tx_headroom = sizeof(struct wfx_hif_msg) + sizeof(struct wfx_hif_req_tx) +
> > 4 /* alignment */ + 8 /* TKIP IV */;
>
> Do you think the fix should rather be:
>
> ------8<----------8<--------
> --- i/net/mac80211/rc80211_minstrel_ht.c
> +++ w/net/mac80211/rc80211_minstrel_ht.c
> @@ -1559,7 +1559,7 @@ minstrel_ht_update_rates(struct minstrel_priv *mp, struct minstrel_ht_sta *mi)
> minstrel_ht_set_rate(mp, mi, rates, i++, mi->max_tp_rate[0]);
>
> /* Fill up remaining, keep one entry for max_probe_rate */
> - for (; i < (mp->hw->max_rates - 1); i++)
> + for (; i < min(mp->hw->max_rates, IEEE80211_TX_RATE_TABLE_SIZE) - 1; i++)
> minstrel_ht_set_rate(mp, mi, rates, i, mi->max_tp_rate[i]);
>
> if (i < mp->hw->max_rates)
> ------8<----------8<--------
>
We thought about this as well - or about adding assertion to the function which does the memory allocation,
but there are more 4-element arrays in mac80211, handled under different defines, which can be confusing.
carl9170 driver has BUILD_BUG_ON to guard from that precisely - see:
https://elixir.bootlin.com/linux/latest/source/drivers/net/wireless/ath/carl9170/tx.c#L879
I think, that the second BUILD_BUG_ON could be moved to mac80211 core, so that it is checked always,
not only when CARL9170 is enabled.
I think both changes should be applied - or, alternatively, in minstrel_ht_set_rate, we could use:
BUG_ON(mp->hw->max_rates > IEEE80211_TX_RATE_TABLE_SIZE);
to quickly catch misbehaving drivers in future.
Since this concerns mac80211 core, let's add Johannes to the loop as well.
>
>
> --
> Jérôme Pouiller
>
>
--
Pozdrawiam/With kind regards,
Lech Perczak
Sr. Software Engineer
Camlin Technologies Poland Limited Sp. z o.o.
Strzegomska 54,
53-611 Wroclaw
Tel: (+48) 71 75 000 16
Email: lech.perczak@xxxxxxxxxxxxxxx
Website: http://www.camlingroup.com
