Because it caused a chroot ttyname regression in 2.6.36.
As of 2.6.36 ttyname does not work in a chroot. It has already been
reported that screen breaks, and for me this breaks an automated
distribution testsuite, that I need to preserve the ability to run the
existing binaries on for several more years. glibc 2.11.3 which has a
fix for this is not an option.
The root cause of this breakage is:
commit 8df9d1a414
Author: Miklos Szeredi <[email protected]>
Date: Tue Aug 10 11:41:41 2010 +0200
vfs: show unreachable paths in getcwd and proc
Prepend "(unreachable)" to path strings if the path is not reachable
from the current root.
Two places updated are
- the return string from getcwd()
- and symlinks under /proc/$PID.
Other uses of d_path() are left unchanged (we know that some old
software crashes if /proc/mounts is changed).
Signed-off-by: Miklos Szeredi <[email protected]>
Signed-off-by: Al Viro <[email protected]>
So remove the nice sounding, but ultimately ill advised change to how
/proc/fd symlinks work.
Signed-off-by: "Eric W. Biederman" <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This fixes a regression in 2.6.35 from 2.6.34, that is
present for select models of Intel cpus when people are
using an MP table.
The commit cf7500c0ea
"x86, ioapic: In mpparse use mp_register_ioapic" started
calling mp_register_ioapic from MP_ioapic_info. An extremely
simple change that was obviously correct. Unfortunately
mp_register_ioapic did just a little more than the previous
hand crafted code and so we gained this call path.
The problem call path is:
MP_ioapic_info()
mp_register_ioapic()
io_apic_unique_id()
io_apic_get_unique_id()
get_physical_broadcast()
modern_apic()
lapic_get_version()
apic_read(APIC_LVR)
Which turned out to be a problem because the local apic
was not mapped, at that point, unlike the similar point
in the ACPI parsing code.
This problem is fixed by mapping the local apic when
parsing the mptable as soon as we reasonably can.
Looking at the number of places we setup the fixmap for
the local apic, I see some serious simplification opportunities.
For the moment except for not duplicating the setting up of the
fixmap in init_apic_mappings, I have not acted on them.
The regression from 2.6.34 is tracked in bug
https://bugzilla.kernel.org/show_bug.cgi?id=16173
Cc: <[email protected]> 2.6.35
Reported-by: David Hill <[email protected]>
Reported-by: Tvrtko Ursulin <[email protected]>
Tested-by: Tvrtko Ursulin <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
Supporting symlinks from untagged to tagged directories is reasonable,
and needed to support CONFIG_SYSFS_DEPRECATED. So don't fail a prior
allowing that case to work.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This happens for network devices when SYSFS_DEPRECATED is enabled.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Recently my tagged sysfs support revealed a flaw in the device core
that a few rare drivers are running into such that we don't always put
network devices in a class subdirectory named net/.
Since we are not creating the class directory the network devices wind
up in a non-tagged directory, but the symlinks to the network devices
from /sys/class/net are in a tagged directory. All of which works
until we go to remove or rename the symlink. When we remove or rename
a symlink we look in the namespace of the target of the symlink.
Since the target of the symlink is in a non-tagged sysfs directory we
don't have a namespace to look in, and we fail to remove the symlink.
Detect this problem up front and simply don't create symlinks we won't
be able to remove later. This prevents symlink leakage and fails in
a much clearer and more understandable way.
Signed-off-by: Eric W. Biederman <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Rafael J. Wysocki <[email protected]>
Cc: Maciej W. Rozycki <[email protected]>
Cc: Kay Sievers <[email protected]>
Cc: Johannes Berg <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This fixes the regression in 2.6.35-rcX where bluetooth network devices
would fail to be deleted from sysfs, causing their destruction and
recreation to fail. In addition this fixes the mac80211_hwsim driver
where it would leave around sysfs files when the driver was removed.
This problem is discussed at
https://bugzilla.kernel.org/show_bug.cgi?id=16257
The reason for the regression is that the network namespace support
added to sysfs expects and requires that network devices be put in
directories that can contain only network devices.
Today get_device_parent almost provides that guarantee for all class
devices, except for a specific exception when the parent of a class
devices is a class device. It would be nice to simply remove that
arguably incorrect special case, but apparently the input devices depend
on it being there. So I have only removed it for class devices with
network namespace support. Which today are the network devices.
It has been suggested that a better fix would be to change the parent
device from a class device to a bus device, which in the case of the
bluetooth driver would change /sys/class/bluetooth to /sys/bus/bluetoth,
I can not see how we would avoid significant userspace breakage if we
were to make that change.
Adding an extra directory in the path to the device will also be
userspace visible but it is much less likely to break things.
Everything is still accessible from /sys/class (for example), and it
fixes two bugs. Adding an extra directory fixes a 3 year old regression
introduced with the new sysfs layout that makes it impossible to rename
bnep0 network devices to names that conflict with hci device attributes
like hci_revsion. Adding an additional directory removes the new
failure modes introduced by the network namespace code.
If it weren't for the regession in the renaming of network devices I
would figure out how to just make the sysfs code deal with this
configuration of devices.
In summary this patch fixes regressions by changing:
"/sys/class/bluetooth/hci0/bnep0" to "/sys/class/bluetooth/hci0/net/bnep0".
Reported-by: Johannes Berg <[email protected]>
Reported-by: Janusz Krzysztofik <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Remove the restriction that only allows connecting to a unix domain
socket identified by unix path that is in the same network namespace.
Crossing network namespaces is always tricky and we did not support
this at first, because of a strict policy of don't mix the namespaces.
Later after Pavel proposed this we did not support this because no one
had performed the audit to make certain using unix domain sockets
across namespaces is safe.
What fundamentally makes connecting to af_unix sockets in other
namespaces is safe is that you have to have the proper permissions on
the unix domain socket inode that lives in the filesystem. If you
want strict isolation you just don't create inodes where unfriendlys
can get at them, or with permissions that allow unfriendlys to open
them. All nicely handled for us by the mount namespace and other
standard file system facilities.
I looked through unix domain sockets and they are a very controlled
environment so none of the work that goes on in dev_forward_skb to
make crossing namespaces safe appears needed, we are not loosing
controll of the skb and so do not need to set up the skb to look like
it is comming in fresh from the outside world. Further the fields in
struct unix_skb_parms should not have any problems crossing network
namespaces.
Now that we handle SCM_CREDENTIALS in a way that gives useable values
across namespaces. There does not appear to be any operational
problems with encouraging the use of unix domain sockets across
containers either.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: Daniel Lezcano <[email protected]>
Acked-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
In unix_skb_parms store pointers to struct pid and struct cred instead
of raw uid, gid, and pid values, then translate the credentials on
reception into values that are meaningful in the receiving processes
namespaces.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Start capturing not only the userspace pid, uid and gid values of the
sending process but also the struct pid and struct cred of the sending
process as well.
This is in preparation for properly supporting SCM_CREDENTIALS for
sockets that have different uid and/or pid namespaces at the different
ends.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: Serge E. Hallyn <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
scm_send occasionally allocates state in the scm_cookie, so I have
modified netlink_sendmsg to guarantee that when scm_send succeeds
scm_destory will be called to free that state.
Signed-off-by: Eric W. Biederman <[email protected]>
Reviewed-by: Daniel Lezcano <[email protected]>
Acked-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Use struct pid and struct cred to store the peer credentials on struct
sock. This gives enough information to convert the peer credential
information to a value relative to whatever namespace the socket is in
at the time.
This removes nasty surprises when using SO_PEERCRED on socket
connetions where the processes on either side are in different pid and
user namespaces.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: Daniel Lezcano <[email protected]>
Acked-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
To keep the coming code clear and to allow both the sock
code and the scm code to share the logic introduce a
fuction to translate from struct cred to struct ucred.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Define what happens when a we view a uid from one user_namespace
in another user_namepece.
- If the user namespaces are the same no mapping is necessary.
- For most cases of difference use overflowuid and overflowgid,
the uid and gid currently used for 16bit apis when we have a 32bit uid
that does fit in 16bits. Effectively the situation is the same,
we want to return a uid or gid that is not assigned to any user.
- For the case when we happen to be mapping the uid or gid of the
creator of the target user namespace use uid 0 and gid as confusing
that user with root is not a problem.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: Serge E. Hallyn <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Reorder the fields in scm_cookie so they pack better on 64bit.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
When I introduced the global variable gsi_end I thought gsi_end on
io_apics was one past the end of the gsi range for the io_apic. After
it was pointed out the the range on io_apics was inclusive I changed
my global variable to match. That was a big mistake. Inclusive
semantics without a range start cannot describe the case when no gsi's
are allocated. Describing the case where no gsi's are allocated is
important in sfi.c and mpparse.c so that we can assign gsi numbers
instead of blindly copying the gsi assignments the BIOS has done as we
do in the acpi case.
To keep from getting the global variable confused with the gsi range
end rename it gsi_top.
To allow describing the case where no gsi's are allocated have gsi_top
be one place the highest gsi number seen in the system.
This fixes an off by one bug in sfi.c:
Reported-by: jacob pan <[email protected]>
This fixes the same off by one bug in mpparse.c:
This fixes an off unreachable by one bug in acpi/boot.c:irq_to_gsi
Reported-by: Yinghai <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
This reverts commit aaf8cdc34d.
Drivers like the ipw2100 call device_create_group when they
are initialized and device_remove_group when they are shutdown.
Moving them between namespaces deletes their sysfs groups early.
In particular the following call chain results.
netdev_unregister_kobject -> device_del -> kobject_del -> sysfs_remove_dir
With sysfs_remove_dir recursively deleting all of it's subdirectories,
and nothing adding them back.
Ouch!
Therefore we need to call something that ultimate calls sysfs_mv_dir
as that sysfs function can move sysfs directories between namespaces
without deleting their subdirectories or their contents. Allowing
us to avoid placing extra boiler plate into every driver that does
something interesting with sysfs.
Currently the function that provides that capability is device_rename.
That is the code works without nasty side effects as originally written.
So remove the misguided fix for moving devices between namespaces. The
bug in the kobject layer that inspired it has now been recognized and
fixed.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: David S. Miller <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
It only makes sense for uevent_helper to get events
in the intial namespaces. It's invocation is not
per namespace and it is not clear how we could make
it's invocation namespace aware.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: David S. Miller <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Utilize netlink_broacast_filtered to allow sending hotplug events
in the proper namespace.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: David S. Miller <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
When netlink sockets are used to convey data that is in a namespace
we need a way to select a subset of the listening sockets to deliver
the packet to. For the network namespace we have been doing this
by only transmitting packets in the correct network namespace.
For data belonging to other namespaces netlink_bradcast_filtered
provides a mechanism that allows us to examine the destination
socket and to decide if we should transmit the specified packet
to it.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: David S. Miller <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
I had a couple of stupid bugs in:
netns: Teach network device kobjects which namespace they are in.
- I duplicated the Kconfig for the NET_NS
- The build was broken when sysfs was not compiled in
The sysfs breakage is because after I moved the operations
for the sysfs to the kobject layer, to make things cleaner
I forgot to move the ifdefs. Opps.
I'm not quite certain how I got introduced a second NET_NS Kconfig,
but it was probably a 3 way merge somewhere along the way that
did not notice that the NET_NS Kconfig option had mvoed and thout
that was a bug. It probably slipped in because it used to be the
sysfs patches were the first patches in my network namespace patches.
Some things just don't go like you would expect.
Neither of these bugs actually affect anything in the common case
but they should be fixed.
Thanks to Serge for noticing they were present.
Reported-by: Serge E. Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: David S. Miller <[email protected]>
The problem. Network devices show up in sysfs and with the network
namespace active multiple devices with the same name can show up in
the same directory, ouch!
To avoid that problem and allow existing applications in network namespaces
to see the same interface that is currently presented in sysfs, this
patch enables the tagging directory support in sysfs.
By using the network namespace pointers as tags to separate out the
the sysfs directory entries we ensure that we don't have conflicts
in the directories and applications only see a limited set of
the network devices.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: David S. Miller <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Open a copy of the uevent kernel socket in each network
namespace so we can send uevents in all network namespaces.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: David S. Miller <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The problem. When implementing a network namespace I need to be able
to have multiple network devices with the same name. Currently this
is a problem for /sys/class/net/*, /sys/devices/virtual/net/*, and
potentially a few other directories of the form /sys/ ... /net/*.
What this patch does is to add an additional tag field to the
sysfs dirent structure. For directories that should show different
contents depending on the context such as /sys/class/net/, and
/sys/devices/virtual/net/ this tag field is used to specify the
context in which those directories should be visible. Effectively
this is the same as creating multiple distinct directories with
the same name but internally to sysfs the result is nicer.
I am calling the concept of a single directory that looks like multiple
directories all at the same path in the filesystem tagged directories.
For the networking namespace the set of directories whose contents I need
to filter with tags can depend on the presence or absence of hotplug
hardware or which modules are currently loaded. Which means I need
a simple race free way to setup those directories as tagged.
To achieve a reace free design all tagged directories are created
and managed by sysfs itself.
Users of this interface:
- define a type in the sysfs_tag_type enumeration.
- call sysfs_register_ns_types with the type and it's operations
- sysfs_exit_ns when an individual tag is no longer valid
- Implement mount_ns() which returns the ns of the calling process
so we can attach it to a sysfs superblock.
- Implement ktype.namespace() which returns the ns of a syfs kobject.
Everything else is left up to sysfs and the driver layer.
For the network namespace mount_ns and namespace() are essentially
one line functions, and look to remain that.
Tags are currently represented a const void * pointers as that is
both generic, prevides enough information for equality comparisons,
and is trivial to create for current users, as it is just the
existing namespace pointer.
The work needed in sysfs is more extensive. At each directory
or symlink creating I need to check if the directory it is being
created in is a tagged directory and if so generate the appropriate
tag to place on the sysfs_dirent. Likewise at each symlink or
directory removal I need to check if the sysfs directory it is
being removed from is a tagged directory and if so figure out
which tag goes along with the name I am deleting.
Currently only directories which hold kobjects, and
symlinks are supported. There is not enough information
in the current file attribute interfaces to give us anything
to discriminate on which makes it useless, and there are
no potential users which makes it an uninteresting problem
to solve.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Benjamin Thery <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Move complete knowledge of namespaces into the kobject layer
so we can use that information when reporting kobjects to
userspace.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
It appears gcc can't cope with using an enum that is only declared in
an inline function declaration, that doesn't even use the variable
that is so declared.
Avoid the silliness and replace the enum with an int, and make gcc
happy.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: Randy Dunlap <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
device_del and device_rename were modified to use
sysfs_delete_link and sysfs_rename_link respectively to ensure
when these operations happen on devices whose classes
are in namespace directories they work properly.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Benjamin Thery <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
When removing a symlink sysfs_remove_link does not provide
enough information to figure out which tagged directory the symlink
falls in. So I need sysfs_delete_link which is passed the target
of the symlink to delete.
sysfs_rename_link is updated to call sysfs_delete_link instead
of sysfs_remove_link as we have all of the information necessary
and the callers are interesting.
Both of these functions now have enough information to find a symlink
in a tagged directory. The only restriction is that they must be called
before the target kobject is renamed or deleted. If they are called
later I loose track of which tag the target kobject was marked with
and can no longer find the old symlink to remove it.
This patch was split from an earlier patch.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Benjamin Thery <[email protected]>
Signed-off-by: Daniel Lezcano <[email protected]>
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Add all of the necessary bioler plate to support
multiple superblocks in sysfs.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
My recent changes introducing a global gsi_end variable
failed to take into account the case of using acpi on a system
not built to support IO_APICs, causing the build to fail.
Define gsi_end to 15 when CONFIG_X86_IO_APIC is not set to avoid
compile errors.
Signed-off-by: Eric W. Biederman <[email protected]>
Cc: Yinghai Lu <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
Now that the generic irq layer is performing the exact same remapping as
io_apic_renumber_irq we can kill this weird es7000 specific function.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
ACPI irq source overrides are allowed for the 16 isa irqs and are
allowed to map any gsi to any isa irq. A few motherboards have been
seen to take advantage of this and put the isa irqs on the 2nd or
3rd ioapic. This causes some problems, most notably the fact
that we can not use any gsi < 16.
To correct this move the gsis that are not isa irqs and have
a gsi number < 16 into the linux irq space just past gsi_end.
This is what the es7000 platform is doing today. Moving only the
low 16 gsis above the rest of the gsi's only penalizes weird
platforms, leaving sane acpi implementations with a 1-1 mapping
of gsis and irqs.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
Use the global gsi_end value now that all ioapics have
valid gsi numbers instead of a combination of acpi_probe_gsi
and walking all of the ioapics and couting their number of
entries by hand if acpi_probe_gsi gave us an answer we did
not like.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
Now that all ioapics have valid gsi_base values use this to
accellerate pin_2_irq. In the case of acpi this also ensures
that pin_2_irq will compute the same irq value for an ioapic
pin as acpi will.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
Now that all ioapic registration happens in mp_register_ioapic we can
move the calculation of nr_ioapic_registers there from enable_IO_APIC.
The number of ioapic registers is already calucated in mp_register_ioapic
so all that really needs to be done is to save the caluclated value
in nr_ioapic_registers.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
Long ago MP_ioapic_info was the primary way of setting up our
ioapic data structures and mp_register_ioapic was a compatibility
shim for acpi code. Now the situation is reversed and
and mp_register_ioapic is the primary way of setting up our
ioapic data structures.
Keep the setting up of ioapic data structures uniform by
having mp_register_ioapic call mp_register_ioapic.
This changes a few fields:
- type: is now hardset to MP_IOAPIC but type had to
bey MP_IOAPIC or MP_ioapic_info would not have been called.
- flags: is now hard coded to MPC_APIC_USABLE.
We require flags to contain at least MPC_APIC_USEBLE in
MP_ioapic_info and we don't ever examine flags so dropping
a few flags that might possibly exist that we have never
used is harmless.
- apicaddr: Unchanged
- apicver: Read from the ioapic instead of using the cached
hardware value in the MP table. The real hardware value
will be more accurate.
- apicid: Now verified to be unique and changed if it is not.
If the BIOS got this right this is a noop. If the BIOS did
not fixing things appears to be the better solution.
This adds gsi_base and gsi_end values to our ioapics defined with
the mpatable, which will make our lives simpler later since
we can always assume gsi_base and gsi_end are valid.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
Add the global variable gsi_end and teach mp_register_ioapic
to keep it uptodate as we add more ioapics into the system.
ioapics can only be added early in boot so the code that
runs later can treat gsi_end as a constant.
Remove the have hacks in sfi.c to second guess mp_register_ioapic
by keeping t's own running total of how many gsi's have been seen,
and instead use the gsi_end.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
This patches fixes the types of gsi_base and gsi_end values in
struct mp_ioapic_gsi, and the gsi parameter of mp_find_ioapic
and mp_find_ioapic_pin
A gsi is cannonically a u32, not an int.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
io_apic_redir_entries has a huge conceptual bug. It returns the maximum
redirection entry not the number of redirection entries. Which simply
does not match what the name of the function. This just caught me
and it caught Feng Tang, and Len Brown when they wrote sfi_parse_ioapic.
Modify io_apic_redir_entries to actually return the number of redirection
entries, and fix the callers so that they properly handle receiving the
number of the number of redirection table entries, instead of the
number of redirection table entries less one.
While the usage in sfi.c does not show up in this patch it is fixed
by virtue of the fact that io_apic_redir_entries now has the semantics
sfi_parse_ioapic most reasonably expects.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
Multiple declarations of the same function in different headers
is a pain to maintain.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
Remove the assumption that there is not an override for isa irq 0.
Instead lookup the gsi and from that lookup the ioapic and pin of each
isa irq indivdually.
In general this should not have any behavioural affect but in
perverse cases this gets all of the details correct, instead of
doing something weird.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
Currently acpi_sci_ioapic_setup calls mp_override_legacy_irq with
bus_irq == gsi, which is wrong if we are comming from an override
Instead pass the bus_irq into acpi_sci_ioapic_setup.
This fix was inspired by a similar fix from:
Yinghai Lu <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
Strictly speaking on x86 (where acpi is used) dev->irq must be
a dual i8259 irq input aka an isa irq. Therefore we should translate
that isa irq into a gsi before passing it to a function that
takes a gsi.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
In perverse acpi implementations the isa irqs are not identity mapped
to the first 16 gsi. Furthermore at least the extended interrupt
resource capability may return gsi's and not isa irqs. So since
what we get from acpi is a gsi teach acpi_get_overrride_irq to
operate on a gsi instead of an isa_irq.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
There are a number of cases where the current code makes the assumption
that isa irqs identity map to the first 16 acpi global system intereupts.
In most instances that assumption is correct as that is the required
behaviour in dual i8259 mode and the default behavior in ioapic mode.
However there are some systems out there that take advantage of acpis
interrupt remapping for the isa irqs to have a completely different
mapping of isa_irq to gsi.
Introduce acpi_isa_irq_to_gsi to perform this mapping explicitly in the
code that needs it. Initially this will be just the current assumed
identity mapping to ensure it's introduction does not cause regressions.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
I haven't been able to get link up on a NX_P3_B1 since 2.6.31. The
driver complains about a firmware hang instead. When I asked I was
told rev 0x41 was a preproduction rev. So disable support in the
driver so no one is surprised the code doesn't work.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Currently sysfs_get_inode magically returns an inode on
sysfs_sb. Make the super_block parameter explicit and
the code becomes clearer.
Acked-by: Tejun Heo <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Don't open code the renaming of symlinks in sysfs
instead use the new helper function sysfs_rename_link
Acked-by: Tejun Heo <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Because of rename ordering problems we occassionally give false
warnings about invalid sysfs operations. So using sysfs_rename
create a sysfs_rename_link function that doesn't need strange
workarounds.
Cc: Benjamin Thery <[email protected]>
Cc: Daniel Lezcano <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
A little more whack-a-mole annotating the dynamic sysfs attributes. I
had everything built into my earlier test kernel, and so I missed
these.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
I have added a new requirement to the external sysfs interface
that dynamically allocated sysfs attributes must call sysfs_attr_init
if lockdep is enabled. For the time being callying sysfs_attr_init
is only mandatory if lockdep is enabled, so we can live with a few
unconverted instances until we find them all. As this is part of
the public interface of sysfs it is a good idea to document these
pseudo functions so someone inspeciting the code can find out
what has happened.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
These are the non-static sysfs attributes that exist on
my test machine. Fix them to use sysfs_attr_init or
sysfs_bin_attr_init as appropriate. It simply requires
making a sysfs attribute present to see this. So this
is a little bit tedious but otherwise not too bad.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: WANG Cong <[email protected]>
Cc: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Acknowledge that the logical sysfs rwsem has one instance per
sysfs attribute with different locking depencencies for different
attributes.
There is a sysfs idiom where writing to one sysfs file causes the
addition or removal of other sysfs files. Lumping all of the
sysfs attributes together in one lock class causes lockdep to
generate lots of false positives.
This introduces the requirement that non-static sysfs attributes
need to be initialized with sysfs_attr_init or sysfs_bin_attr_init.
Strictly speaking this requirement only exists when lockdep is
enabled, and when lockdep is enabled we get a bit fat warning
if this requirement is not met.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: WANG Cong <[email protected]>
Cc: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
If we exclude directories and symlinks from the set of sysfs
dirents where we need active references we are left with
sysfs attributes (binary or not).
- Tweak sysfs_deactivate to only do something on attributes
- Move lockdep initialization into sysfs_file_add_mode to
limit it to just attributes.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: WANG Cong <[email protected]>
Cc: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
It turns out that holding an active reference on a directory is
pointless. The purpose of the active references are to allows us to
block when removing sysfs entries that have custom methods so we don't
remove modules while running modular code and to keep those custom
methods from accessing data structures after the files have been
removed. Further sysfs_remove_dir remove all elements in the
directory before removing the directory itself, so there is no chance
we will remove a directory with active children.
Signed-off-by: Eric W. Biederman <[email protected]>
Cc: Tejun Heo <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
When sysfs_readdir stops short we now cache the next
sysfs_dirent to return to user space in filp->private_data.
There is no impact on the rest of sysfs by doing this and
in the common case it allows us to pick up exactly where
we left off with no seeking.
Additionally I drop and regrab the sysfs_mutex around
filldir to avoid a page fault abritrarily increasing the
hold time on the sysfs_mutex.
v2: Returned to using INT_MAX as the EOF condition.
seekdir is ambiguous unless all directory entries have
a unique f_pos value.
Fixes http://bugzilla.kernel.org/show_bug.cgi?id=14949
Signed-off-by: Eric W. Biederman <[email protected]>
Cc: Linus Torvalds <[email protected]>
Cc: stable <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
To avoid potential problems with an empty /dev open /dev/console
from rootfs instead of waiting to mount our root filesystem and
mounting it there. This effectively guarantees that there will
be a device node, and it won't be on a filesystem that we will
ever unmount, so there are no issues with leaving /dev/console
open and pinning the filesystem.
This is actually more effective than automatically mounting
devtmpfs on /dev because it removes removes the occasionally
problematic assumption that /dev/console exists from the boot
code.
With this patch I was able to throw busybox on my /boot partition
(which has no /dev directory) and boot into userspace without
problems.
The only possible negative consequence I can think of is that
someone out there deliberately used did not use a character device
that is major 5 minor 2 for /dev/console. Does anyone know of a
situation in which that could make sense?
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Al Viro <[email protected]>
We use scm_send and scm_recv on both unix domain and
netlink sockets, but only unix domain sockets support
everything required for file descriptor passing,
so error if someone attempts to pass file descriptors
over netlink sockets.
Cc: [email protected]
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Iranna D Ankad reported that IBM x3950 systems have boot
problems after this commit:
|
| commit b9c61b7007
|
| x86/pci: update pirq_enable_irq() to setup io apic routing
|
The problem is that with the patch, the machine freezes when
console=ttyS0,... kernel serial parameter is passed.
It seem to freeze at DVD initialization and the whole problem
seem to be DVD/pata related, but somehow exposed through the
serial parameter.
Such apic problems can expose really weird behavior:
ACPI: IOAPIC (id[0x10] address[0xfecff000] gsi_base[0])
IOAPIC[0]: apic_id 16, version 0, address 0xfecff000, GSI 0-2
ACPI: IOAPIC (id[0x0f] address[0xfec00000] gsi_base[3])
IOAPIC[1]: apic_id 15, version 0, address 0xfec00000, GSI 3-38
ACPI: IOAPIC (id[0x0e] address[0xfec01000] gsi_base[39])
IOAPIC[2]: apic_id 14, version 0, address 0xfec01000, GSI 39-74
ACPI: INT_SRC_OVR (bus 0 bus_irq 1 global_irq 4 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 5 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 3 global_irq 6 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 4 global_irq 7 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 6 global_irq 9 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 7 global_irq 10 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 8 global_irq 11 low edge)
ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 12 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 12 global_irq 15 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 13 global_irq 16 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 14 global_irq 17 low edge)
ACPI: INT_SRC_OVR (bus 0 bus_irq 15 global_irq 18 dfl dfl)
It turns out that the system has three io apic controllers, but
boot ioapic routing is in the second one, and that gsi_base is
not 0 - it is using a bunch of INT_SRC_OVR...
So these recent changes:
1. one set routing for first io apic controller
2. assume irq = gsi
... will break that system.
So try to remap those gsis, need to seperate boot_ioapic_idx
detection out of enable_IO_APIC() and call them early.
So introduce boot_ioapic_idx, and remap_ioapic_gsi()...
-v2: shift gsi with delta instead of gsi_base of boot_ioapic_idx
-v3: double check with find_isa_irq_apic(0, mp_INT) to get right
boot_ioapic_idx
-v4: nr_legacy_irqs
-v5: add print out for boot_ioapic_idx, and also make it could be
applied for current kernel and previous kernel
-v6: add bus_irq, in acpi_sci_ioapic_setup, so can get overwride
for sci right mapping...
-v7: looks like pnpacpi get irq instead of gsi, so need to revert
them back...
-v8: split into two patches
-v9: according to Eric, use fixed 16 for shifting instead of remap
-v10: still need to touch rsparser.c
-v11: just revert back to way Eric suggest...
anyway the ioapic in first ioapic is blocked by second...
-v12: two patches, this one will add more loop but check apic_id and irq > 16
Reported-by: Iranna D Ankad <[email protected]>
Bisected-by: Iranna D Ankad <[email protected]>
Tested-by: Gary Hade <[email protected]>
Signed-off-by: Yinghai Lu <[email protected]>
Cc: Eric W. Biederman <[email protected]>
Cc: Thomas Renninger <[email protected]>
Cc: Eric W. Biederman <[email protected]>
Cc: Suresh Siddha <[email protected]>
Cc: [email protected]
LKML-Reference: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
The wireless sysfs methods like the rest of the networking sysfs
methods are removed with the rtnl_lock held and block until
the existing methods stop executing. So use rtnl_trylock
and restart_syscall so that the code continues to work.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Yuck. It turns out that when we restart sysctls we were restarting
with the values already changed. Which unfortunately meant that
the second time through we thought there was no change and skipped
all kinds of work, despite the fact that there was indeed a change.
I have fixed this the simplest way possible by restoring the changed
values when we restart the sysctl write.
One of my coworkers spotted this bug when after disabling forwarding
on an interface pings were still forwarded.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Right now xen's use of the x86 and ia64 handle_irq is just bizarre and very
fragile as it is very non-obvious the function exists and is is used by
code out in drivers/.... Luckily using handle_irq is completely unnecessary,
and we can just use the generic irq apis instead.
This still leaves drivers/xen/events.c as a problematic user of the generic
irq apis it has "static struct irq_info irq_info[NR_IRQS]" but that can be
fixed some other time.
Signed-off-by: Eric W. Biederman <[email protected]>
LKML-Reference: <[email protected]>
Acked-by: Jeremy Fitzhardinge <[email protected]>
Cc: Ian Campbell <[email protected]>
Signed-off-by: H. Peter Anvin <[email protected]>
Stop computing the number of neighbour table settings we have by
counting the number of binary sysctls. This behaviour was silly
and meant that we could not add another neighbour table setting
without also adding another binary sysctl.
Don't pass the binary sysctl path for neighour table entries
into neigh_sysctl_register. These parameters are no longer
used and so are just dead code.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Stop using the binary sysctl enumeartion in sysctl.h as an index into
a per interface array. This leads to unnecessary binary sysctl number
allocation, and a fragility in data structure and implementation
because of unnecessary coupling.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
There is currently a bug in sysfs_sd_setattr inherited from
sysfs_setattr in 2.6.32 where the first time we set the attributes
on a sysfs file we allocate backing store but do not set the
backing store attributes. Resulting in overly restrictive
permissions on sysfs files.
The fix is to simply modify the code so that it always executes
when we update the sysfs attributes, as we did in 2.6.31 and earlier.
Signed-off-by: Eric W. Biederman <[email protected]>
Tested-by: Jean Delvare <[email protected]>
Cc: stable <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
sysfs_remove_group() waits for sysfs attributes to be removed, therefore
we do not need to worry about driver-specific attributes being accessed
after driver has been detached from the device. In fact, attempts to take
serio->drv_mutex in attribute methods may lead to the following deadlock:
sysfs_read_file()
fill_read_buffer()
sysfs_get_active_two()
psmouse_attr_show_helper()
serio_pin_driver()
serio_disconnect_driver()
mutex_lock(&serio->drv_mutex);
<--------> mutex_lock(&serio_drv_mutex);
psmouse_disconnect()
sysfs_remove_group(... psmouse_attr_group);
....
sysfs_deactivate();
wait_for_completion();
Fix this by removing calls to serio_[un]pin_driver() and functions themselves
and using driver-private mutexes to serialize access to attribute's set()
methods that may change device state.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Dmitry Torokhov <[email protected]>
Holding locks over device_del -> kobject_del -> sysfs_deactivate can
cause deadlocks if those same locks are grabbed in sysfs show or store
methods.
The I model s_active count + completion as a sleeping read/write lock.
I describe to lockdep sysfs_get_active as a read_trylock,
sysfs_put_active as a read_unlock, and sysfs_deactivate as a
write_lock and write_unlock pair. This seems to capture the essence
for purposes of finding deadlocks, and in my testing gives finds real
issues and ignores non-issues.
This brings us back to holding locks over kobject_del is a problem
that ideally we should find a way of addressing, but at least lockdep
can tell us about the problems instead of requiring developers to debug
rare strange system deadlocks, that happen when sysfs files are removed
while being written to.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
I received some bug reports about userspace programs having problems
because after RTM_NEWLINK was received they could not immediate access
files under /proc/sys/net/ because they had not been registered yet.
The original problem was trivially fixed by moving the userspace
notification from rtnetlink_event() to the end of
register_netdevice().
When testing that change I discovered I was still getting RTM_NEWLINK
events before I could access proc and I was also getting RTM_NEWLINK
events after I was seeing RTM_DELLINK. Things practically guaranteed
to confuse userspace.
After a little more investigation these extra notifications proved to
be from the new notifiers NETDEV_POST_INIT and NETDEV_UNREGISTER_BATCH
hitting the default case in rtnetlink_event, and triggering
unnecessary RTM_NEWLINK messages.
rtnetlink_event now explicitly handles NETDEV_UNREGISTER_BATCH and
NETDEV_POST_INIT to avoid sending the incorrect userspace
notifications.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
These two functions do 90% of the same work and it doesn't significantly
obfuscate the function to allow both the parent dir and the name to change
at the same time. So merge them together to simplify maintenance, and
increase testing.
Acked-by: Tejun Heo <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
By teaching sysfs_revalidate to hide a dentry for
a sysfs_dirent if the sysfs_dirent has been renamed,
and by teaching sysfs_lookup to return the original
dentry if the sysfs dirent has been renamed. I can
show the results of renames correctly without having to
update the dcache during the directory rename.
This massively simplifies the rename logic allowing a lot
of weird sysfs special cases to be removed along with
a lot of now unnecesary helper code.
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Now that sysfs_getattr and sysfs_permission refresh the vfs
inode there is no need to immediatly push the mode change
into the vfs cache. Reducing the amount of work needed and
simplifying the locking.
Acked-by: Tejun Heo <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
With the implementation of sysfs_getattr and sysfs_permission
sysfs becomes able to lazily propogate inode attribute changes
from the sysfs_dirents to the vfs inodes. This paves the way
for deleting significant chunks of now unnecessary code.
While doing this we did not reference sysfs_setattr from
sysfs_symlink_inode_operations so I added along with
sysfs_getattr and sysfs_permission.
Acked-by: Tejun Heo <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Lining up the functions in sysfs_symlink_inode_operations
follows the pattern in the rest of sysfs and makes things
slightly more readable.
Acked-by: Tejun Heo <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Currently sysfs updates the timestamps on the vfs directory
inode when we create or remove a directory entry but doesn't
update the cached copy on the sysfs_dirent, fix that oversight.
Acked-by: Tejun Heo <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Cleanly separate the work that is specific to setting the
attributes of a sysfs_dirent from what is needed to update
the attributes of a vfs inode.
Additionally grab the sysfs_mutex to keep any nasties from
surprising us when updating the sysfs_dirent.
Acked-by: Tejun Heo <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The granularity of sysfs time when we keep it is 1 ns. Which
when passed to timestamp_trunc results in a nop. So remove
the unnecessary function call making sysfs_setattr slightly
easier to read.
Acked-by: Tejun Heo <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Currently every caller of sysfs_chmod_file happens at either
file creation time to set a non-default mode or in response
to a specific user requested space change in policy. Making
timestamps of when the chmod happens and notification of
a file changing mode uninteresting.
Remove the unnecessary time stamp and filesystem change
notification, and removes the last of the explicit inotify
and donitfy support from sysfs.
Acked-by: Tejun Heo <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Calling d_drop unconditionally when a sysfs_dirent is deleted has
the potential to leak mounts, so instead implement dentry delete
and revalidate operations that cause sysfs dentries to be removed
at the appropriate time.
Acked-by: Tejun Heo <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Using dentry instead of d in the function name is what
several other filesystems are doing and it seems to be
a more readable convention.
Acked-by: Tejun Heo <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
device_shutdown is defined to just shutdown the hardware and to not
clean up any kernel data structures. Therefore don't put the kobjects
for /sys/dev and /sys/dev/block and /sys/dev/char.
This ensures we don't remove /sys/dev/block and /sys/dev/char while
we still have symlinks from there to the actual devices.
Acked-by: Kay Sievers <[email protected]>
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This function walks the whole hashtable so there is no point in
passing it a network namespace. Instead I purge all timewait
sockets from dead network namespaces that I find. If the namespace
is one of the once I am trying to purge I am guaranteed no new timewait
sockets can be formed so this will get them all. If the namespace
is one I am not acting for it might form a few more but I will
call inet_twsk_purge again and shortly to get rid of them. In
any even if the network namespace is dead timewait sockets are
useless.
Move the calls of inet_twsk_purge into batch_exit routines so
that if I am killing a bunch of namespaces at once I will just
call inet_twsk_purge once and save a lot of redundant unnecessary
work.
My simple 4k network namespace exit test the cleanup time dropped from
roughly 8.2s to 1.6s. While the time spent running inet_twsk_purge fell
to about 2ms. 1ms for ipv4 and 1ms for ipv6.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
While we are looking up entries to free there is no reason to take
the lock in inet_twsk_purge. We have to drop locks and restart
occassionally anyway so adding a few more in case we get on the
wrong list because of a timewait move is no big deal. At the
same time not taking the lock for long periods of time is much
more polite to the rest of the users of the hash table.
In my test configuration of killing 4k network namespaces
this change causes 4k back to back runs of inet_twsk_purge on an
empty hash table to go from roughly 20.7s to 3.3s, and the total
time to destroy 4k network namespaces goes from roughly 44s to
3.3s.
Signed-off-by: Eric W. Biederman <[email protected]>
Acked-by: Eric Dumazet <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Refactor the code so fib_rules_register always takes a template instead
of the actual fib_rules_ops structure that will be used. This is
required for network namespace support so 2 out of the 3 callers already
do this, it allows the error handling to be made common, and it allows
fib_rules_unregister to free the template for hte caller.
Modify fib_rules_unregister to use call_rcu instead of syncrhonize_rcu
to allw multiple namespaces to be cleaned up in the same rcu grace
period.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This allows namespace exit methods to batch work that comes requires an
rcu barrier using call_rcu without having to treat the
unregister_pernet_operations cases specially.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
xfrm.nlsk is provided by the xfrm_user module and is access via rcu from
other parts of the xfrm code. Add xfrm.nlsk_stash a copy of xfrm.nlsk that
will never be set to NULL. This allows the synchronize_net and
netlink_kernel_release to be deferred until a whole batch of xfrm.nlsk sockets
have been set to NULL.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Move network device exit batching from a special case in
net_namespace.c to using common mechanisms in dev.c
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
- Add exit_list to struct net to support building lists of network
namespaces to cleanup.
- Add exit_batch to pernet_operations to allow running operations only
once during a network namespace exit. Instead of once per network
namespace.
- Factor opt ops_exit_list and ops_exit_free so the logic with cleanup
up a network namespace does not need to be duplicated.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Take advantage of the fact that an explicit rtnl_kill_links is
unnecessary (and skipping it improves batching), as network namespace
exit calls dellink on all remaining virtual devices, and
rtnl_link_unregister calls dellink on all outstanding devices in that
network namespace. To do this we need to leave the vlan proc
directories in place until after network device exit time, which is
done by using register_pernet_subsys instead of
register_pernet_device.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
No that all of the callers have been updated to set fields in
struct pernet_operations, and simplified to let the network
namespace core handle the allocation and freeing of the storage
for them, remove the surpurpflous methods and update the docs
to the new style.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Take advantage of the new pernet automatic storage management,
and stop using compatibility network namespace functions.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Take advantage of the new pernet automatic storage management,
and stop using compatibility network namespace functions.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Take advantage of the new pernet automatic storage management,
and stop using compatibility network namespace functions.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Take advantage of the new pernet automatic storage management,
and stop using compatibility network namespace functions.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Take advantage of the new pernet automatic storage management,
and stop using compatibility network namespace functions.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Take advantage of the new pernet automatic storage management,
and stop using compatibility network namespace functions.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Take advantage of the new pernet automatic storage management,
and stop using compatibility network namespace functions.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Take advantage of the new pernet automatic storage management,
and stop using compatibility network namespace functions.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Take advantage of the new pernet automatic storage management,
and stop using compatibility network namespace functions.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Take advantage of the new pernet automatic storage management,
and stop using compatibility network namespace functions.
Signed-off-by: Eric W. Biederman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Eric W. Biederman