systemd/1001-re-enable-by_path-links-for-ata-devices.patch
Dominique Leuenberger a4023d0c37 Accepting request 333777 from Base:System
- Fix patch tty-ask-password-agent-on-console.patch not to crash
  away but enable it to ask on all devices of /dev/console

- Avoid "file not found /etc/udev/rules.d/70-persistent-net.rules"
  waring occurring in %post

- Add patch let-vconsole-setup-get-properties-only-once-to-copy-them.patch
  to avoid broken virtual console mapping due stressed ioctl API
  for the virtual consoles (boo#904214)

- Fix last change that is use the new name for udev packages in %pretrans. 

- restore usage of LUA in %pretrans.

- Try to generate the systemd users and groups always in same order
  to avoid republish other packages (boo#944660) 

- cleanup specfile by removing commands that were dealing with systemd
  pre-generated files: we're now using systemd tarball generated directly
  from the git repo which doesn't contain any of these files.
- there's no point in using LUA in %pretrans

- Drop 0009-make-xsltproc-use-correct-ROFF-links.patch
This patch was initialy added to workaround bsc#842844. But it
appears that man(1) was fixed (included since 13.2 at least) to
handle manual pages that consist only of a .so directive such as
'.so <page>'.

- Change use-rndaddentropy-ioctl-to-load-random-seed.patch to
  make it work on big endian

OBS-URL: https://build.opensuse.org/request/show/333777
OBS-URL: https://build.opensuse.org/package/show/openSUSE:Factory/systemd?expand=0&rev=223
2015-09-27 12:32:31 +00:00

100 lines
3.5 KiB
Diff

From 5cf46aa4339670afac386b1b0e630b739f3621c7 Mon Sep 17 00:00:00 2001
From: Robert Milasan <rmilasan@suse.com>
Date: Thu, 12 Jul 2012 15:56:34 +0000
Subject: [PATCH] Persistent by_path links for ata devices
With newer kernel we have the 'port_no' attribute,
which allows us to construct a valid ata by-path link.
With this patch ATA links of the form
ata-<port>.[01]
(for master/slave devices) or
ata-<port>.<pmp>.0
(for devices behind port multipliers)
are generated.
References: bnc#770910,FATE#317063
Signed-off-by: Robert Milasan <rmilasan@suse.com>
Signed-off-by: Hannes Reinecke <hare@suse.de>
---
src/udev/udev-builtin-path_id.c | 53 ++++++++++++++++++++++++++++++----------
1 file changed, 41 insertions(+), 12 deletions(-)
Index: systemd-221/src/udev/udev-builtin-path_id.c
===================================================================
--- systemd-221.orig/src/udev/udev-builtin-path_id.c
+++ systemd-221/src/udev/udev-builtin-path_id.c
@@ -447,6 +447,46 @@ static struct udev_device *handle_scsi_h
return parent;
}
+static struct udev_device *handle_ata(struct udev_device *parent, char **path)
+{
+ struct udev *udev = udev_device_get_udev(parent);
+ struct udev_device *hostdev, *portdev;
+ int host, bus, target, lun, port_no;
+ const char *name, *atahost, *port;
+
+ hostdev = udev_device_get_parent_with_subsystem_devtype(parent, "scsi", "scsi_host");
+ if (hostdev == NULL)
+ return NULL;
+
+ name = udev_device_get_sysname(parent);
+ if (sscanf(name, "%d:%d:%d:%d", &host, &bus, &target, &lun) != 4)
+ return NULL;
+
+ /* The ata port is the parent of the SCSI host */
+ hostdev = udev_device_get_parent(hostdev);
+ atahost = udev_device_get_sysname(hostdev);
+ if (strncmp(atahost, "ata", 3))
+ return NULL;
+
+ /* ATA port number is found in 'port_no' attribute */
+ portdev = udev_device_new_from_subsystem_sysname(udev, "ata_port",
+ atahost);
+ port = udev_device_get_sysattr_value(portdev, "port_no");
+ if (!port || sscanf(port, "%d", &port_no) != 1) {
+ hostdev = NULL;
+ goto out;
+ }
+ if (bus != 0)
+ /* Devices behind port multiplier have a bus != 0*/
+ path_prepend(path, "ata-%u.%u.0", port_no, bus);
+ else
+ /* Master/slave are distinguished by target id */
+ path_prepend(path, "ata-%u.%u", port_no, target);
+out:
+ udev_device_unref(portdev);
+ return hostdev;
+}
+
static struct udev_device *handle_scsi(struct udev_device *parent, char **path, bool *supported_parent) {
const char *devtype;
const char *name;
@@ -486,19 +526,8 @@ static struct udev_device *handle_scsi(s
goto out;
}
- /*
- * We do not support the ATA transport class, it uses global counters
- * to name the ata devices which numbers spread across multiple
- * controllers.
- *
- * The real link numbers are not exported. Also, possible chains of ports
- * behind port multipliers cannot be composed that way.
- *
- * Until all that is solved at the kernel level, there are no by-path/
- * links for ATA devices.
- */
if (strstr(name, "/ata") != NULL) {
- parent = NULL;
+ parent = handle_ata(parent, path);
goto out;
}