* Fix FSID of rootfs being 0

For st_dev in struct stat, the value 0 is
  an error. Make sure rootfs has value 1.

  This fixes the coreutils cp -ax issue (although that
  code has already been improved in coreutils, too) and
  somewhat fixes the switch_root problem in util-linux.
  It is still possible that rootfs gets an FSID >1 in
  future versions, so util-linux still needs to be fixed.

  This is unrelated to the mount id issue in
  /proc/*/mountinfo - the mount id may still be
  zero.

* Revert a patch that always enables autosuspend on intel bluetoothh

  xhci still sucks and breaks this. I am reverting this
  commit for reasons that are in no way selfish.
This commit is contained in:
Thomas Bächler 2014-04-03 22:57:31 +00:00
parent 1dc3c46ebf
commit d068bb28c8
7 changed files with 99 additions and 9 deletions

View File

@ -1,7 +1,7 @@
From 5d77ba2d26110c678b40fd723866a17d4036de12 Mon Sep 17 00:00:00 2001
From: Lucas De Marchi <lucas.demarchi@intel.com>
Date: Tue, 18 Feb 2014 02:19:26 -0300
Subject: [PATCH 1/4] Bluetooth: allocate static minor for vhci
Subject: [PATCH 1/6] Bluetooth: allocate static minor for vhci
Commit bfacbb9 (Bluetooth: Use devname:vhci module alias for virtual HCI
driver) added the module alias to hci_vhci module so it's possible to

View File

@ -1,7 +1,7 @@
From 9bc5b710f5957763d6944f38143b627d127c15ff Mon Sep 17 00:00:00 2001
From: Tom Gundersen <teg@jklm.no>
Date: Mon, 3 Feb 2014 11:14:13 +1030
Subject: [PATCH 2/4] module: allow multiple calls to MODULE_DEVICE_TABLE() per
Subject: [PATCH 2/6] module: allow multiple calls to MODULE_DEVICE_TABLE() per
module
Commit 78551277e4df5: "Input: i8042 - add PNP modaliases" had a bug, where the

View File

@ -1,7 +1,7 @@
From 06d2e746733a83469944481cb7f4fb1c7134a8ce Mon Sep 17 00:00:00 2001
From: Rusty Russell <rusty@rustcorp.com.au>
Date: Mon, 3 Feb 2014 11:15:13 +1030
Subject: [PATCH 3/4] module: remove MODULE_GENERIC_TABLE
Subject: [PATCH 3/6] module: remove MODULE_GENERIC_TABLE
MODULE_DEVICE_TABLE() calles MODULE_GENERIC_TABLE(); make it do the
work directly. This also removes a wart introduced in the last patch,

View File

@ -1,7 +1,7 @@
From 268984331a319f47179d6f0b62aa0e312fe4497d Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Thomas=20B=C3=A4chler?= <thomas@archlinux.org>
Date: Tue, 1 Apr 2014 01:15:34 +0200
Subject: [PATCH 4/4] Revert "syscalls.h: use gcc alias instead of assembler
Subject: [PATCH 4/6] Revert "syscalls.h: use gcc alias instead of assembler
aliases for syscalls"
This reverts commit 83460ec8dcac14142e7860a01fa59c267ac4657c.

View File

@ -0,0 +1,44 @@
From 0de4ffcad031c1d0b8422ebcceb9caa80995a73e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Thomas=20B=C3=A4chler?= <thomas@archlinux.org>
Date: Thu, 3 Apr 2014 21:55:37 +0200
Subject: [PATCH 5/6] fs: Don't return 0 from get_anon_bdev
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Commit 9e30cc9595303b27b48 removed an internal mount. This
has the side-effect that rootfs now has FSID 0. Many
userspace utilities assume that st_dev in struct stat
is never 0, so this change breaks a number of tools in
early userspace.
Since we don't know how many userspace programs are affected,
make sure that FSID is at least 1.
References: http://article.gmane.org/gmane.linux.kernel/1666905
References: http://permalink.gmane.org/gmane.linux.utilities.util-linux-ng/8557
Cc: 3.14 <stable@vger.kernel.org>
Signed-off-by: Thomas Bächler <thomas@archlinux.org>
---
fs/super.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/fs/super.c b/fs/super.c
index 80d5cf2..7624267 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -802,7 +802,10 @@ void emergency_remount(void)
static DEFINE_IDA(unnamed_dev_ida);
static DEFINE_SPINLOCK(unnamed_dev_lock);/* protects the above */
-static int unnamed_dev_start = 0; /* don't bother trying below it */
+/* Many userspace utilities consider an FSID of 0 invalid.
+ * Always return at least 1 from get_anon_bdev.
+ */
+static int unnamed_dev_start = 1;
int get_anon_bdev(dev_t *p)
{
--
1.9.1

View File

@ -0,0 +1,33 @@
From d02e0fc02f3cfa078e301698f0475adcdf6f72ab Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Thomas=20B=C3=A4chler?= <thomas@archlinux.org>
Date: Thu, 3 Apr 2014 23:59:49 +0200
Subject: [PATCH 6/6] Revert "Bluetooth: Enable autosuspend for Intel Bluetooth
device"
This reverts commit d2bee8fb6e18f6116aada39851918473761f7ab1.
USB autosuspend still breaks on some xhci controllers, so disable
it by default as long as no solution is found.
---
drivers/bluetooth/btusb.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
index baeaaed..6d6e09e 100644
--- a/drivers/bluetooth/btusb.c
+++ b/drivers/bluetooth/btusb.c
@@ -1478,10 +1478,8 @@ static int btusb_probe(struct usb_interface *intf,
if (id->driver_info & BTUSB_BCM92035)
hdev->setup = btusb_setup_bcm92035;
- if (id->driver_info & BTUSB_INTEL) {
- usb_enable_autosuspend(data->udev);
+ if (id->driver_info & BTUSB_INTEL)
hdev->setup = btusb_setup_intel;
- }
/* Interface numbers are hardcoded in the specification */
data->isoc = usb_ifnum_to_if(data->udev, 1);
--
1.9.1

View File

@ -5,7 +5,7 @@ pkgbase=linux # Build stock -ARCH kernel
#pkgbase=linux-custom # Build kernel with a different name
_srcname=linux-3.14
pkgver=3.14
pkgrel=2
pkgrel=3
arch=('i686' 'x86_64')
url="http://www.kernel.org/"
license=('GPL2')
@ -22,16 +22,20 @@ source=("https://www.kernel.org/pub/linux/kernel/v3.x/${_srcname}.tar.xz"
'0002-module-allow-multiple-calls-to-MODULE_DEVICE_TABLE-p.patch'
'0003-module-remove-MODULE_GENERIC_TABLE.patch'
'0004-Revert-syscalls.h-use-gcc-alias-instead-of-assembler.patch'
'0005-fs-Don-t-return-0-from-get_anon_bdev.patch'
'0006-Revert-Bluetooth-Enable-autosuspend-for-Intel-Blueto.patch'
)
sha256sums=('61558aa490855f42b6340d1a1596be47454909629327c49a5e4e10268065dffa'
'91b05561c96e09c8c3a4ea05fdab0745f578c4bb879feae25f552306bbe6522a'
'00da25459826be1dced644c2215dfe2a6da53168d92aea79467c490aec4cbcae'
'f0d90e756f14533ee67afda280500511a62465b4f76adcc5effa95a40045179c'
'faced4eb4c47c4eb1a9ee8a5bf8a7c4b49d6b4d78efbe426e410730e6267d182'
'29a958c37e1fb04c49b1cc3be4f38b53ffdeef61ab049865209f6b77817302d0'
'd8db3b235d871d2b27986d3cb73a57b1d33fed52196a4b608152498a57e93838'
'eb9a773b7b978bc508e8459ef835ff2cb8a41bb235caff1c18c5b718d7ce902c'
'4edd92642d710d8a351dd2acce36d412eccbf60380f9d0368ec315acb49ad704')
'ad22f6e1a2869730828ddaf93da8b8a748421b7afdd34e1213566985fe7b01dd'
'29e612a8dda2b2b8af116c16b546a0e1d159f249bfd9628d640cad0f61f0cc4f'
'555133445de48e781ced3a450bcdf9bfbe4d66b118bbd5b28fea789fa023dd66'
'dc1eca93dad93c9c015332d188f5a69508c2077536b8be653d2ff45f1b034941'
'1f572e748d96403218e846ec469474f557f04d2f5ff137dd2a31fffa193f760e'
'a413700adf3b45b7d3c18d2ffed390b9e5769cb361a9a0530f98bfff29fd0aef')
_kernelname=${pkgbase#linux}
@ -60,6 +64,15 @@ prepare() {
# For details, see https://lkml.org/lkml/2014/1/26/22
patch -p1 -i "${srcdir}/0004-Revert-syscalls.h-use-gcc-alias-instead-of-assembler.patch"
# Fix various bugs caused by rootfs having FSID 0
# See http://www.spinics.net/lists/kernel/msg1716924.html
patch -p1 -i "${srcdir}/0005-fs-Don-t-return-0-from-get_anon_bdev.patch"
# Disable usb autosuspend for intel btusb
# See http://www.spinics.net/lists/kernel/msg1716461.html
# Until a solution is found, make sure the driver leaves autosuspend alone
patch -p1 -i "${srcdir}/0006-Revert-Bluetooth-Enable-autosuspend-for-Intel-Blueto.patch"
if [ "${CARCH}" = "x86_64" ]; then
cat "${srcdir}/config.x86_64" > ./.config
else