PCI devices like QCA6390 load the board file differently, add support for that
and the method is chosen using bus_params variables.
Add support to create board name for different targets. This board name is
used to parse the board data from board-2.bin for ahb/pci based targets.
As struct target_mem_chunk::vaddr was changed from 'u32' to 'u32 *' in
ath11k_qmi_assign_target_mem_chunk() vaddr assignments were changed to NULL to
avoid a compilation warning. IPQ8074 does not use the vaddr field for anything
so that change does not affect functionality.
At the moment this only supports board files with BIN type. Support for ELF
type, which seems to be more popular on QCA6390 devices, needs to be added later.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.1.0.1-01238-QCAHKSWPL_SILICONZ-2
Signed-off-by: Govind Singh <govinds@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1597389030-13887-3-git-send-email-kvalo@codeaurora.org
Modem Host Interface (MHI) is a communication protocol to communicate with
external Qualcomm modems and Wi-Fi chipsets over high speed peripheral buses.
Even though MHI doesn’t dictate underlying physical layer, protocol and MHI
stack is structured for PCI based devices.
Register directly with MHI subsystem as a MHI device driver for firmware
download to QCA6390.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.1.0.1-01238-QCAHKSWPL_SILICONZ-2
Signed-off-by: Govind Singh <govinds@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1597309466-19688-9-git-send-email-kvalo@codeaurora.org
This is needed to initialise hw_params before MHI registration starts. MHI
needs location of firmware directory and that's delivered via hw_params.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.1.0.1-01238-QCAHKSWPL_SILICONZ-2
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1597309466-19688-8-git-send-email-kvalo@codeaurora.org
With only IPQ8074 supported ath11k was only usable on Qualcomm architectures.
But now that we are adding QCA6390 PCI support to ath11k that's not the case
anymore and it can be used on any architecture supporting PCI. So remove the
dependency on ARCH_QCOM. After that there is also no need to depend on
COMPILE_TEST.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.1.0.1-01238-QCAHKSWPL_SILICONZ-2
Signed-off-by: Carl Huang <cjhuang@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1597309466-19688-2-git-send-email-kvalo@codeaurora.org
The mcs index of HT mode is 0 to 31, please refer http://mcsindex.com/.
Its spatial stream(Nss) number is from 1 to 4, mcs index is 0~7 for
Nss=1, 8~15 for Nss=2, 16~23 for Nss=3 and 24~31 is for Nss=4.
The mcs is reported from firmware in wmi_tlv_peer_stats_info of
event WMI_TLV_PEER_STATS_INFO_EVENTID, its range is from 0~15 for
QCA6174 SDIO and PCIe. It is for both Nss=1 and Nss=2, and it has 2
rate table supported_ht_mcs_rate_nss1 and supported_ht_mcs_rate_nss2
in ath10k, they are for Nss=1 and Nss=2, each table has 8 rates.
It need to find the matched row number with the mcs index, for example,
mcs index is 2, it is <=7, so it is Nss=1, and match row 2 in table of
Nss=1. If mcs index is 12, it is >= 8 and <= 15, so it is Nss=2, it
match row 4(12-8) in table of Nss=2. If mcs index is >=16, it is for
Nss=3/4, it need to add rate table, so it is not support in current
ath10k.
This patch is to find the row number in rate table of Nss=1 or Nss=2
with the mcs index reported from firmware.
This patch only effect the chips which supports_peer_stats_info of its
hw_params is true, it is true only for QCA6174 currently.
Tested-on: QCA6174 hw3.2 SDIO WLAN.RMH.4.4.1-00048
Tested-on: QCA6174 hw3.2 PCI WLAN.RM.4.4.1-00110-QCARMSWP-1
Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1597392294-13124-1-git-send-email-wgong@codeaurora.org
When using QCA6174 PCI devices working in station mode, after connected
to AP, tx bitrate is always '1.0 MBit/s' in output of command 'iw wlan0
station dump'. (QCA6174 SDIO devices are working fine.)
After this patch, it show correct bitrate:
Station c4:04:15:5d:97:22 (on wls1)
inactive time: 312 ms
rx bytes: 31496
rx packets: 173
tx bytes: 8625
tx packets: 46
tx retries: 0
tx failed: 0
signal: -76 [-88, -80] dBm
signal avg: -75 [-82, -77] dBm
tx bitrate: 39.0 MBit/s MCS 4
rx bitrate: 26.0 MBit/s MCS 3
Tested-on: QCA6174 hw3.2 PCI WLAN.RM.4.4.1-00110-QCARMSWP-1
Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1597312029-32348-4-git-send-email-wgong@codeaurora.org
ath10k_sta_statistics is used to report many info to iw wlan0 link,
if it return for empty legacy and nss of arsta->txrate, then the other
stats after it will not be set.
It has 4 bit to set after the return:
NL80211_STA_INFO_TX_FAILED
NL80211_STA_INFO_RX_BITRATE
NL80211_STA_INFO_TX_BITRATE
NL80211_STA_INFO_TX_RETRIES
This patch not effect the info of above 4 bit for all hardware, reason
as below:
NL80211_STA_INFO_TX_FAILED is only for htt.disable_tx_comp is true, it
is for QCA6174 SDIO.
NL80211_STA_INFO_RX_BITRATE and NL80211_STA_INFO_TX_BITRATE are both
set in ath10k_mac_sta_get_peer_stats_info, it is only enabled for chips
which supports_peer_stats_info is true in hw_params, recently only for
QCA6174 SDIO, NL80211_STA_INFO_TX_BITRATE is set again in function
ath10k_mac_sta_get_peer_stats_info because the value which parsed from
arsta->tx_rate_code and arsta->tx_bitrate_kbps is correct for QCA6174
SDIO and PCIe, and the value arsta->txrate is not correct for QCA6174
SDIO and PCIe, so it need to set again with the correct value.
NL80211_STA_INFO_TX_RETRIES is use value of arsta->tx_retries, it is set
in ath10k_update_per_peer_tx_stats, which accumulate the retry_pkts in
HTT message from firmware, if the chips not support this feature, then
it is always 0 after accumulate, then iw wlan0 station dump always show
0 for retry count. If not set NL80211_STA_INFO_TX_RETRIES here, then it
is still 0, so the result is same, then set NL80211_STA_INFO_TX_RETRIES
has no effect.
Tested-on: QCA6174 hw3.2 SDIO WLAN.RMH.4.4.1-00048
Tested-on: QCA6174 hw3.2 PCI WLAN.RM.4.4.1-00110-QCARMSWP-1
Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1597312029-32348-3-git-send-email-wgong@codeaurora.org
ath10k_sta_statistics is used to report info for iw wlan0 link,
it check ath10k_peer_stats_enabled, and ath10k_peer_stats_enabled
check WMI_SERVICE_PEER_STATS bit of ar->wmi.svc_map. SVCMAP() for
WMI_SERVICE_PEER_STATS was defined only for wmi_10x_svc_map and
wmi_10_4_svc_map interfaces, it missed in wmi_tlv_svc_map, so it is
not usable for iw wlan0 link for wmi tlv interface.
If firmware report WMI_TLV_SERVICE_PEER_STATS_INFO for wmi tlv, then
enable the WMI_SERVICE_PEER_STATS bit in ath10k, and then it pass check
in ath10k_peer_stats_enabled and ath10k_sta_statistics pass check.
Tested-on: QCA6174 hw3.2 SDIO WLAN.RMH.4.4.1-00048
Tested-on: QCA6174 hw3.2 PCI WLAN.RM.4.4.1-00110-QCARMSWP-1
Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1597312029-32348-2-git-send-email-wgong@codeaurora.org
When simulate random transfer fail for sdio write and read, it happened
"payload length exceeds max htc length" and recovery later sometimes.
Test steps:
1. Add config and update kernel:
CONFIG_FAIL_MMC_REQUEST=y
CONFIG_FAULT_INJECTION=y
CONFIG_FAULT_INJECTION_DEBUG_FS=y
2. Run simulate fail:
cd /sys/kernel/debug/mmc1/fail_mmc_request
echo 10 > probability
echo 10 > times # repeat until hitting issues
3. It happened payload length exceeds max htc length.
[ 199.935506] ath10k_sdio mmc1:0001:1: payload length 57005 exceeds max htc length: 4088
....
[ 264.990191] ath10k_sdio mmc1:0001:1: payload length 57005 exceeds max htc length: 4088
4. after some time, such as 60 seconds, it start recovery which triggered
by wmi command timeout for periodic scan.
[ 269.229232] ieee80211 phy0: Hardware restart was requested
[ 269.734693] ath10k_sdio mmc1:0001:1: device successfully recovered
The simulate fail of sdio is not a real sdio transter fail, it only
set an error status in mmc_should_fail_request after the transfer end,
actually the transfer is success, then sdio_io_rw_ext_helper will
return error status and stop transfer the left data. For example,
the really RX len is 286 bytes, then it will split to 2 blocks in
sdio_io_rw_ext_helper, one is 256 bytes, left is 30 bytes, if the
first 256 bytes get an error status by mmc_should_fail_request,then
the left 30 bytes will not read in this RX operation. Then when the
next RX arrive, the left 30 bytes will be considered as the header
of the read, the top 4 bytes of the 30 bytes will be considered as
lookaheads, but actually the 4 bytes is not the lookaheads, so the len
from this lookaheads is not correct, it exceeds max htc length 4088
sometimes. When happened exceeds, the buffer chain is not matched between
firmware and ath10k, then it need to start recovery ASAP. Recently then
recovery will be started by wmi command timeout, but it will be long time
later, for example, it is 60+ seconds later from the periodic scan, if
it does not have periodic scan, it will be longer.
Start recovery when it happened "payload length exceeds max htc length"
will be reasonable.
This patch only effect sdio chips.
Tested with QCA6174 SDIO with firmware WLAN.RMH.4.4.1-00029.
Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200108031957.22308-3-wgong@codeaurora.org
Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.
Deterministic algorithm:
For each file:
If not .svg:
For each line:
If doesn't contain `\bxmlns\b`:
For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
If both the HTTP and HTTPS versions
return 200 OK and serve the same content:
Replace HTTP with HTTPS.
Signed-off-by: Alexander A. Klimov <grandmaster@al2klimov.de>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200719104041.57916-1-grandmaster@al2klimov.de
list_for_each_entry{_safe} is able to handle an empty list.
The only effect of avoiding the loop is not initializing the
index variable.
Drop list_empty tests in cases where these variables are not
used.
Note that list_for_each_entry{_safe} is defined in terms of
list_first_entry, which indicates that it should not be used on an
empty list. But in list_for_each_entry{_safe}, the element obtained
by list_first_entry is not really accessed, only the address of its
list_head field is compared to the address of the list head, so the
list_first_entry is safe.
The semantic patch that makes this change for the list_for_each_entry
case is as follows: (http://coccinelle.lip6.fr/)
<smpl>
@@
expression x,e;
statement S;
identifier i;
@@
-if (!(list_empty(x)))
list_for_each_entry(i,x,...) S
... when != i
? i = e
</smpl>
Signed-off-by: Julia Lawall <Julia.Lawall@inria.fr>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1595761112-11003-8-git-send-email-Julia.Lawall@inria.fr
The controller is capable of reporting TX indication which can be used
to report TX ack when IEEE80211_TX_CTL_REQ_TX_STATUS is set.
The support was only partially implemented.
The firmware can be configured for reporting event when a packet is
acked, without specifying which packet though. In order to send a
packet flagged with TX status callback, we need to stop the queue,
submit the packet and wait for the firmware ack event. Then the queue
can be restarted and mac80211 status callback called.
In case the packet is not acked, no ack event will be received,
therefore a timeout mechanism is introduced to restart the queue
and call the status cb in case no event is received after a 100ms.
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1595586052-16081-3-git-send-email-loic.poulain@linaro.org
Several AMPDU sessions can be started, e.g. for different TIDs.
Currently the driver does not take care of the session ID when
requesting block-ack (statically set to 0), which leads to never
block-acked packet with sessions other than 0.
Fix this by saving the session id when creating the ba session and
use it in subsequent ba operations.
This issue can be reproduced with iperf in two steps (tid 0 strem
then tid 6 stream).
1.0 iperf -s # wcn36xx side
1.1 iperf -c ${IP_ADDR} # host side
Then
2.0 iperf -s -u -S 0xC0 # wcn36xx side
2.1 iperf -c ${IP_ADDR} -u -S 0xC0 -l 2000 # host side
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1595586052-16081-2-git-send-email-loic.poulain@linaro.org
Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.
Deterministic algorithm:
For each file:
If not .svg:
For each line:
If doesn't contain `\bxmlns\b`:
For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
If both the HTTP and HTTPS versions
return 200 OK and serve the same content:
Replace HTTP with HTTPS.
Signed-off-by: Alexander A. Klimov <grandmaster@al2klimov.de>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200719105052.57997-1-grandmaster@al2klimov.de
The idx in __ath10k_htt_rx_ring_fill_n function lives in
consistent dma region writable by the device. Malfunctional
or malicious device could manipulate such idx to have a OOB
write. Either by
htt->rx_ring.netbufs_ring[idx] = skb;
or by
ath10k_htt_set_paddrs_ring(htt, paddr, idx);
The idx can also be negative as it's signed, giving a large
memory space to write to.
It's possibly exploitable by corruptting a legit pointer with
a skb pointer. And then fill skb with payload as rougue object.
Part of the log here. Sometimes it appears as UAF when writing
to a freed memory by chance.
[ 15.594376] BUG: unable to handle page fault for address: ffff887f5c1804f0
[ 15.595483] #PF: supervisor write access in kernel mode
[ 15.596250] #PF: error_code(0x0002) - not-present page
[ 15.597013] PGD 0 P4D 0
[ 15.597395] Oops: 0002 [#1] SMP KASAN PTI
[ 15.597967] CPU: 0 PID: 82 Comm: kworker/u2:2 Not tainted 5.6.0 #69
[ 15.598843] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
BIOS rel-1.12.1-0-ga5cab58e9a3f-prebuilt.qemu.org 04/01/2014
[ 15.600438] Workqueue: ath10k_wq ath10k_core_register_work [ath10k_core]
[ 15.601389] RIP: 0010:__ath10k_htt_rx_ring_fill_n
(linux/drivers/net/wireless/ath/ath10k/htt_rx.c:173) ath10k_core
Signed-off-by: Zekun Shen <bruceshenzk@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200623221105.3486-1-bruceshenzk@gmail.com
On failure pcie_capability_read_dword() sets it's last parameter, val
to 0. However, with Patch 14/14, it is possible that val is set to ~0 on
failure. This would introduce a bug because (x & x) == (~0 & x).
This bug can be avoided without changing the function's behaviour if the
return value of pcie_capability_read_dword is checked to confirm success.
Check the return value of pcie_capability_read_dword() to ensure success.
Suggested-by: Bjorn Helgaas <bjorn@helgaas.com>
Signed-off-by: Bolarinwa Olayemi Saheed <refactormyself@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200713175529.29715-2-refactormyself@gmail.com
The downside of using defines in struct ath11k_hw_params.fw.dir is that it's
easy to get it wrong as the full path is not visible. So drop the use of
defines and instead create the patch runtime using a static inline function
ath11k_core_create_firmware_path(). Hopefully this reduces the chances of using
incorrect firmware path.
No functional changes. Compile tested only.
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1592316055-24958-8-git-send-email-kvalo@codeaurora.org
To avoid duplicating the logic how the full firmware path is created create a
common function ath11k_core_firmware_request() and convert also qmi.c to use it.
Also remove a useless info print, it's more like a debug message anyway.
No functional changes. Compile tested only.
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1592316055-24958-7-git-send-email-kvalo@codeaurora.org
A compromized ath10k peripheral is able to control the size argument
of memcpy in ath10k_pci_hif_exchange_bmi_msg.
The min result from previous line is not used as the size argument
for memcpy. Instead, xfer.resp_len comes from untrusted stream dma
input. The value comes from "nbytes" in ath10k_pci_bmi_recv_data,
which is set inside _ath10k_ce_completed_recv_next_nolock with the line
nbytes = __le16_to_cpu(sdesc.nbytes);
sdesc is a stream dma region which device can write to.
Signed-off-by: Zekun Shen <bruceshenzk@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200616132544.17478-1-bruceshenzk@gmail.com
It is expected that the returned counters by .get_survey are monotonic
increasing. But the data from ath10k gets reset to zero regularly. Channel
active/busy time are then showing incorrect values (less than previous or
sometimes zero) for the currently active channel during successive survey
dump commands.
example:
$ iw dev wlan0 survey dump
Survey data from wlan0
frequency: 5180 MHz [in use]
channel active time: 54995 ms
channel busy time: 432 ms
channel receive time: 0 ms
channel transmit time: 59 ms
...
$ iw dev wlan0 survey dump
Survey data from wlan0
frequency: 5180 MHz [in use]
channel active time: 32592 ms
channel busy time: 254 ms
channel receive time: 0 ms
channel transmit time: 0 ms
...
The correct way to handle this is to use the non-clearing
WMI_BSS_SURVEY_REQ_TYPE_READ wmi_bss_survey_req_type. The firmware will
then accumulate the survey data and handle wrap arounds.
Tested-on: QCA9984 hw1.0 10.4-3.5.3-00057
Tested-on: QCA988X hw2.0 10.2.4-1.0-00047
Tested-on: QCA9888 hw2.0 10.4-3.9.0.2-00024
Tested-on: QCA4019 hw1.0 10.4-3.6-00140
Fixes: fa7937e3d5 ("ath10k: update bss channel survey information")
Signed-off-by: Venkateswara Naralasetty <vnaralas@codeaurora.org>
Tested-by: Markus Theil <markus.theil@tu-ilmenau.de>
Tested-by: John Deere <24601deerej@gmail.com>
[sven@narfation.org: adjust commit message]
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1592232686-28712-1-git-send-email-kvalo@codeaurora.org
When tx status enabled, retry count is updated from tx completion status.
which is not working as expected due to firmware limitation where
firmware can not provide per MSDU rate statistics from tx completion
status. Due to this tx retry count is always 0 in station dump.
Fix this issue by updating the retry packet count from per peer
statistics. This patch will not break on SDIO devices since, this retry
count is already updating from peer statistics for SDIO devices.
Tested-on: QCA9984 PCI 10.4-3.6-00104
Tested-on: QCA9882 PCI 10.2.4-1.0-00047
Signed-off-by: Venkateswara Naralasetty <vnaralas@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1591856446-26977-1-git-send-email-vnaralas@codeaurora.org
On wcn3990 we have "per_ce_irq = true". That makes the
ath10k_ce_interrupt_summary() function always return 0xfff. The
ath10k_ce_per_engine_service_any() function will see this and think
that _all_ copy engines have an interrupt. Without checking, the
ath10k_ce_per_engine_service() assumes that if it's called that the
"copy complete" (cc) interrupt fired. This combination seems bad.
Let's add a check to make sure that the "copy complete" interrupt
actually fired in ath10k_ce_per_engine_service().
This might fix a hard-to-reproduce failure where it appears that the
copy complete handlers run before the copy is really complete.
Specifically a symptom was that we were seeing this on a Qualcomm
sc7180 board:
arm-smmu 15000000.iommu: Unhandled context fault:
fsr=0x402, iova=0x7fdd45780, fsynr=0x30003, cbfrsynra=0xc1, cb=10
Even on platforms that don't have wcn3990 this still seems like it
would be a sane thing to do. Specifically the current IRQ handler
comments indicate that there might be other misc interrupt sources
firing that need to be cleared. If one of those sources was the one
that caused the IRQ handler to be called it would also be important to
double-check that the interrupt we cared about actually fired.
Tested-on: WCN3990 SNOC WLAN.HL.3.2.2-00490-QCAHLSWMTPL-1
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200609082015.1.Ife398994e5a0a6830e4d4a16306ef36e0144e7ba@changeid
First set of patches intended for v5.9
* Fix links to wiki;
* Some preparations for gcc-10;
* Make FW reconfiguration quieter by not using warn level;
* Some other small fixes and clean-up;
# gpg: Signature made Thu 11 Jun 2020 12:03:51 PM EEST using RSA key ID 1A3CC5FA
# gpg: Good signature from "Luciano Roth Coelho (Luca) <luca@coelho.fi>"
# gpg: aka "Luciano Roth Coelho (Intel) <luciano.coelho@intel.com>"
