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 for Android: free password hash cracker in your pocket
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20200320231928.137720-1-abhishekpandit@chromium.org>
Date:   Fri, 20 Mar 2020 16:19:27 -0700
From:   Abhishek Pandit-Subedi <abhishekpandit@...omium.org>
To:     marcel@...tmann.org, linux-bluetooth@...r.kernel.org
Cc:     chromeos-bluetooth-upstreaming@...omium.org,
        Abhishek Pandit-Subedi <abhishekpandit@...omium.org>,
        "David S. Miller" <davem@...emloft.net>,
        Johan Hedberg <johan.hedberg@...il.com>,
        netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
        Jakub Kicinski <kuba@...nel.org>
Subject: [PATCH v2 0/1] Bluetooth: Prioritize sco traffic on slow interfaces


Hi Marcel,

While investigating supporting Voice over HCI/UART, we discovered that
it is possible for SCO packet deadlines to be missed in some conditions
where large ACL packets are being transferred. For UART, at a baudrate
of 3000000, a single 1024 byte packet will take ~3.4ms to transfer.
Sending two ACL packets of max size would cause us to miss the timing
for SCO (which is 3.75ms) in the worst case.

To mitigate this, we change hci_tx_work to prefer scheduling SCO/eSCO
over ACL/LE and modify the hci_sched_{acl,le} routines so that they will
only send one packet before checking whether a SCO packet is queued. ACL
packets should still get sent at a similar rate (depending on number of
ACL packets supported by controller) since the loop will continue until
there is no more quota left for ACL and LE packets.

To test this patch, I played some music over SCO (open youtube and
a video conference page at the same time) while using an LE keyboard and
mouse.  There were no discernible slowdowns caused by this change.

Thanks
Abhishek

Changes in v2:
* Refactor to check for SCO/eSCO after each ACL/LE packet sent
* Enabled SCO priority all the time and removed the sched_limit variable

Abhishek Pandit-Subedi (1):
  Bluetooth: Prioritize SCO traffic

 net/bluetooth/hci_core.c | 111 +++++++++++++++++++++------------------
 1 file changed, 61 insertions(+), 50 deletions(-)

-- 
2.25.1.696.g5e7596f4ac-goog

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ