Hi, Recently I upgraded sys-kernel/gentoo-sources from 4.19.37 to 4.19.44 on several stable systems. Today I discovered that the Sony DualShock 3 gamepad is unable to connect via bluetooth when running 4.19.44 kernel. It constantly connect/disconnect with input/output errors. Some other devices (headset) works fine. After some research I found that it seems to be an upstream bug¹ impacting several devices. If I reboot on old 4.19.37 kernel, the gamepad works fine. [1] https://bugzilla.kernel.org/show_bug.cgi?id=203643 Reproducible: Always
This patch¹ applied against gentoo-sources-4.19.44 fixed it for me. 1. https://lore.kernel.org/linux-bluetooth/20190522070540.48895-1-marcel@holtmann.org/T/#u
(In reply to Fab from comment #1) > This patch¹ applied against gentoo-sources-4.19.44 fixed it for me. > > 1. > https://lore.kernel.org/linux-bluetooth/20190522070540.48895-1- > marcel@holtmann.org/T/#u Yes, this Patch fixed here gentoo-sources:4.19.47 for a USB 046d:0b04 Logitech, Inc. Logitech BT Mini-Receiver (for a Logitech diNovo Edge Keyboard) Thank you
Queued for 4.19.50. Thanks for reporting. commit 7918237fb06dafdefd25924598ff1ea91f6dc34a (HEAD -> 4.19, origin/4.19) Author: Mike Pagano <mpagano@gentoo.org> Date: Mon Jun 10 15:41:47 2019 -0400 Bluetooth: Check key sizes only when Secure Simple Pairing is enabled. See bug #686758. Signed-off-by: Mike Pagano <mpagano@gentoo.org>
Released in gentoo-sources-4.19.50 commit 904c16d687359930aecda99df450b05dea839997 (HEAD -> master, origin/master, origin/HEAD) Author: Mike Pagano <mpagano@gentoo.org> Date: Tue Jun 11 10:37:16 2019 -0400 sys-kernel/gentoo-sources: Linux patch 4.19.50 and bluetooth patch Bluetooth: Check key sizes only when Secure Simple Pairing is enabled. See bug #686758.
This bug is also present in newer versions, e.g. gentoo-sources-5.1.7. It should be fixed there as well.
In fact, if you search on kernel.org all branchs which contains the guilty commit d5bb334a8e171b262e48f378bd2096c0ea458265, you will find that the bug is present in following vanilla-sources versions (thus in corresponding gentoo-sources and later versions) : >= 5.1.1 >= 5.0.15 >= 4.19.42 >= 4.14.118 >= 4.9.175 >= 4.4.180 I don't know why the fixed patch take so long to hit the main kernel tree. Is this the usual timelapse between patch proposal and integration in kernel development ?
It seems that the original patch which causes the issue will be reverted in all next longterm/stable releases (from 4.4.x to 5.1.x) : > This reverts commit 38f092c41cebaff589e88cc22686b289a6840559 which is > commit d5bb334a8e171b262e48f378bd2096c0ea458265 upstream. > > Lots of people have reported issues with this patch, and as there does > not seem to be a fix going into Linus's kernel tree any time soon, > revert the commit in the stable trees so as to get people's machines > working properly again. https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/tree/queue-4.19/revert-bluetooth-align-minimum-encryption-key-size-for-le-and-br-edr-connections.patch
Do you have the lkml message for this? I notice the git commit has vanished from the upstream repo. Thanks!
(In reply to Michael 'veremitz' Everitt from comment #8) > Do you have the lkml message for this? I notice the git commit has vanished > from the upstream repo. Thanks! Not vanished from the repo, but from the queue. The revert commit was released into : > 5.1.10 > 4.19.51 ⁽¹⁾ > 4.14.126 I don't know why it is reverted in some longtem kernels, but not reverted in some others. Very strange. 1. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v4.19.52&id=8e5483aeae0df86346f975a06c5a381bc4dd97e7
(In reply to Fab from comment #9) > I don't know why it is reverted in some longtem kernels, but not reverted in > some others. Very strange. > I found the answer : because the revert patch is not yet released for 4.4.x and 4.9.x branchs. For these branchs the revert patch is still in queue-4.4 and queue-4.9 directories⁽¹⁾. So I guess it will be also be released soon. 1. https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/tree/
So, to sum up : The original patch d5bb334a8e171b262e48f378bd2096c0ea458265 which cause the issue was merged into folowing stables/longterms kernels : >= 5.1.1 >= 5.0.15 (EOL) >= 4.19.42 >= 4.14.118 >= 4.9.175 >= 4.4.180 Then, it was reverted a few weeks later into : >= 5.1.10 >= 4.19.51 >= 4.14.126 >= 4.9.183 >= 4.4.183 On the main kernel tree, the original patch was not reverted but an other fix⁽¹⁾ was commited and released into 5.2-rc6. I guess that everything will be backported in stables/longterms trees soon. 1. https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?h=v5.2-rc7&id=693cd8ce3f882524a5d06f7800dd8492411877b3