blueness 14/07/25 12:12:15 Added: eudev-exclude-MD.patch Log: Add upstream patch for excluding md devices from block device ownership, bug #517986 (Portage version: 2.2.8-r1/cvs/Linux x86_64, signed Manifest commit with key 0xF52D4BBA)
Revision Changes Path 1.1 sys-fs/eudev/files/eudev-exclude-MD.patch file : http://sources.gentoo.org/viewvc.cgi/gentoo-x86/sys-fs/eudev/files/eudev-exclude-MD.patch?rev=1.1&view=markup plain: http://sources.gentoo.org/viewvc.cgi/gentoo-x86/sys-fs/eudev/files/eudev-exclude-MD.patch?rev=1.1&content-type=text/plain Index: eudev-exclude-MD.patch =================================================================== >From 2e5b17d01347d3c3118be2b8ad63d20415dbb1f0 Mon Sep 17 00:00:00 2001 From: Kay Sievers <k...@vrfy.org> Date: Thu, 24 Jul 2014 23:37:35 +0200 Subject: [PATCH] udev: exclude MD from block device ownership event locking MD instantiates devices at open(). This is incomptible with the locking logic, as the "change" event emitted when stopping a device will bring it back. --- src/udev/udevd.c | 23 +++++++---------------- 1 file changed, 7 insertions(+), 16 deletions(-) diff --git a/src/udev/udevd.c b/src/udev/udevd.c index a45d324..db935d6 100644 --- a/src/udev/udevd.c +++ b/src/udev/udevd.c @@ -285,26 +285,17 @@ static void worker_new(struct event *event) udev_event->exec_delay = exec_delay; /* - * Take a "read lock" on the device node; this establishes + * Take a shared lock on the device node; this establishes * a concept of device "ownership" to serialize device - * access. External processes holding a "write lock" will + * access. External processes holding an exclusive lock will * cause udev to skip the event handling; in the case udev - * acquired the lock, the external process will block until + * acquired the lock, the external process can block until * udev has finished its event handling. */ - - /* - * <kabi_> since we make check - device seems unused - we try - * ioctl to deactivate - and device is found to be opened - * <kay> sure, you try to take a write lock - * <kay> if you get it udev is out - * <kay> if you can't get it, udev is busy - * <kabi_> we cannot deactivate openned device (as it is in-use) - * <kay> maybe we should just exclude dm from that thing entirely - * <kabi_> IMHO this sounds like a good plan for this moment - */ - if (streq_ptr("block", udev_device_get_subsystem(dev)) && - !startswith(udev_device_get_sysname(dev), "dm-")) { + if (!streq_ptr(udev_device_get_action(dev), "remove") && + streq_ptr("block", udev_device_get_subsystem(dev)) && + !startswith(udev_device_get_sysname(dev), "dm-") && + !startswith(udev_device_get_sysname(dev), "md")) { struct udev_device *d = dev; if (streq_ptr("partition", udev_device_get_devtype(d))) -- 1.8.5.5
