Avi Kivity [Thu, 7 Jun 2007 16:18:30 +0000 (19:18 +0300)]
KVM: Flush remote tlbs when reducing shadow pte permissions
When a vcpu causes a shadow tlb entry to have reduced permissions, it
must also clear the tlb on remote vcpus. We do that by:
- setting a bit on the vcpu that requests a tlb flush before the next entry
- if the vcpu is currently executing, we send an ipi to make sure it
exits before we continue
Avi Kivity [Tue, 5 Jun 2007 09:17:03 +0000 (12:17 +0300)]
KVM: Fix vcpu freeing for guest smp
A vcpu can pin up to four mmu shadow pages, which means the freeing
loop will never terminate. Fix by first unpinning shadow pages on
all vcpus, then freeing shadow pages.
Avi Kivity [Mon, 4 Jun 2007 12:58:30 +0000 (15:58 +0300)]
KVM: Lazy guest cr3 switching
Switch guest paging context may require us to allocate memory, which
might fail. Instead of wiring up error paths everywhere, make context
switching lazy and actually do the switch before the next guest entry,
where we can return an error if allocation fails.
Avi Kivity [Thu, 31 May 2007 15:24:09 +0000 (18:24 +0300)]
KVM: MMU: Don't cache guest access bits in the shadow page table
This was once used to avoid accessing the guest pte when upgrading
the shadow pte from read-only to read-write. But usually we need
to set the guest pte dirty or accessed bits anyway, so this wasn't
really exploited.
Avi Kivity [Thu, 31 May 2007 15:20:14 +0000 (18:20 +0300)]
KVM: MMU: Simpify accessed/dirty/present/nx bit handling
Always set the accessed and dirty bit (since having them cleared causes
a read-modify-write cycle), always set the present bit, and copy the
nx bit from the guest.
Avi Kivity [Thu, 31 May 2007 12:23:35 +0000 (15:23 +0300)]
KVM: Make shadow pte updates atomic
With guest smp, a second vcpu might see partial updates when the first
vcpu services a page fault. So delay all updates until we have figured
out what the pte should look like.
Note that on i386, this is still not completely atomic as a 64-bit write
will be split into two on a 32-bit machine.
Jan Engelhardt [Wed, 23 May 2007 21:22:11 +0000 (14:22 -0700)]
Use menuconfig objects II - KVM/Virt
Make a "menuconfig" out of the Kconfig objects "menu, ..., endmenu",
so that the user can disable all the options in that menu at once
instead of having to disable each option separately.
Signed-off-by: Jan Engelhardt <jengelh@gmx.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Avi Kivity <avi@qumranet.com>
Eddie Dong [Mon, 21 May 2007 04:28:09 +0000 (07:28 +0300)]
KVM: VMX: Avoid saving and restoring msr_efer on lightweight vmexit
MSR_EFER.LME/LMA bits are automatically save/restored by VMX
hardware, KVM only needs to save NX/SCE bits at time of heavy
weight VM Exit. But clearing NX bits in host envirnment may
cause system hang if the host page table is using EXB bits,
thus we leave NX bits as it is. If Host NX=1 and guest NX=0, we
can do guest page table EXB bits check before inserting a shadow
pte (though no guest is expecting to see this kind of gp fault).
If host NX=0, we present guest no Execute-Disable feature to guest,
thus no host NX=0, guest NX=1 combination.
This patch reduces raw vmexit time by ~27%.
Me: fix compile warnings on i386.
Signed-off-by: Yaozu (Eddie) Dong <eddie.dong@intel.com> Signed-off-by: Avi Kivity <avi@qumranet.com>
Eddie Dong [Thu, 17 May 2007 15:55:15 +0000 (18:55 +0300)]
KVM: VMX: Avoid saving and restoring msrs on lightweight vmexit
In a lightweight exit (where we exit and reenter the guest without
scheduling or exiting to userspace in between), we don't need various
msrs on the host, and avoiding shuffling them around reduces raw exit
time by 8%.
i386 compile fix by Daniel Hecken <dh@bahntechnik.de>.
Signed-off-by: Yaozu (Eddie) Dong <eddie.dong@intel.com> Signed-off-by: Avi Kivity <avi@qumranet.com>
Avi Kivity [Mon, 7 May 2007 07:55:37 +0000 (10:55 +0300)]
KVM: Avoid corrupting tr in real mode
The real mode tr needs to be set to a specific tss so that I/O
instructions can function. Divert the new tr values to the real
mode save area from where they will be restored on transition to
protected mode.
This fixes some crashes on reboot when the bios accesses an I/O
instruction.
Avi Kivity [Sun, 6 May 2007 13:10:01 +0000 (16:10 +0300)]
KVM: VMX: Only reload guest msrs if they are already loaded
If we set an msr via an ioctl() instead of by handling a guest exit, we
have the host state loaded, so reloading the msrs would clobber host
state instead of guest state.
This fixes a host oops (and loss of a cpu) on a guest reboot.
Matthew Gregan [Sun, 6 May 2007 07:59:46 +0000 (10:59 +0300)]
KVM: Implement IA32_EBL_CR_POWERON msr
Attempting to boot the default 'bsd' kernel of OpenBSD 4.1 i386 in a guest
fails early in the kernel init inside p3_get_bus_clock while trying to read
the IA32_EBL_CR_POWERON MSR. KVM logs an 'unhandled MSR' message and the
guest kernel faults.
This patch is sufficient to allow OpenBSD to boot, after which it seems to
run fine. I'm not sure if this is the correct solution for dealing with
this particular MSR, but it works for me.
Signed-off-by: Matthew Gregan <kinetik@flim.org> Signed-off-by: Avi Kivity <avi@qumranet.com>
Avi Kivity [Wed, 2 May 2007 14:57:40 +0000 (17:57 +0300)]
KVM: Rationalize exception bitmap usage
Everyone owns a piece of the exception bitmap, but they happily write to
the entire thing like there's no tomorrow. Centralize handling in
update_exception_bitmap() and have everyone call that.
Avi Kivity [Wed, 2 May 2007 13:54:03 +0000 (16:54 +0300)]
KVM: Fix potential guest state leak into host
The lightweight vmexit path avoids saving and reloading certain host
state. However in certain cases lightweight vmexit handling can schedule()
which requires reloading the host state.
So we store the host state in the vcpu structure, and reloaded it if we
relinquish the vcpu.
Avi Kivity [Tue, 1 May 2007 13:53:31 +0000 (16:53 +0300)]
KVM: Update shadow pte on write to guest pte
A typical demand page/copy on write pattern is:
- page fault on vaddr
- kvm propagates fault to guest
- guest handles fault, updates pte
- kvm traps write, clears shadow pte, resumes guest
- guest returns to userspace, re-faults on same vaddr
- kvm installs shadow pte, resumes guest
- guest continues
So, three vmexits for a single guest page fault. But if instead of clearing
the page table entry, we update to correspond to the value that the guest
has just written, we eliminate the third vmexit.
This patch does exactly that, reducing kbuild time by about 10%.
Avi Kivity [Tue, 1 May 2007 13:44:05 +0000 (16:44 +0300)]
KVM: MMU: Respect nonpae pagetable quadrant when zapping ptes
When a guest writes to a page that has an mmu shadow, we have to clear
the shadow pte corresponding to the memory location touched by the guest.
Now, in nonpae mode, a single guest page may have two or four shadow
pages (because a nonpae page maps 4MB or 4GB, whereas the pae shadow maps
2MB or 1GB), so we when we look up the page we find up to three additional
aliases for the page. Since we _clear_ the shadow pte, it doesn't matter
except for a slight performance penalty, but if we want to _update_ the
shadow pte instead of clearing it, it is vital that we don't modify the
aliases.
Fortunately, exactly which page is needed (the "quadrant") is easily
computed, and is accessible in the shadow page header. All we need is
to ignore shadow pages from the wrong quadrants.
Avi Kivity [Tue, 1 May 2007 08:32:28 +0000 (11:32 +0300)]
KVM: Be more careful restoring fs on lightweight vmexit
i386 wants fs for accessing the pda even on a lightweight exit, so ensure
we can always restore it. This fixes a regression on i386 introduced by
the lightweight vmexit patch.
Avi Kivity [Mon, 30 Apr 2007 14:05:38 +0000 (17:05 +0300)]
KVM: Reduce misfirings of the fork detector
The kvm mmu tries to detects forks by looking for repeated writes to a
page table. If it sees a fork, it unshadows the page table so the page
table copying can proceed at native speed instead of being emulated.
However, the detector also triggered on simple demand paging access patterns:
a linear walk of memory would of course cause repeated writes to the same
pagetable page, causing it to unshadow prematurely.
Fix by resetting the fork detector if we detect a demand fault.
Avi Kivity [Mon, 30 Apr 2007 13:07:54 +0000 (16:07 +0300)]
KVM: Avoid saving and restoring some host CPU state on lightweight vmexit
Many msrs and the like will only be used by the host if we schedule() or
return to userspace. Therefore, we avoid saving them if we handle the
exit within the kernel, and if a reschedule is not requested.
Based on a patch from Eddie Dong <eddie.dong@intel.com> with a couple of
fixes by me.
Signed-off-by: Yaozu(Eddie) Dong <eddie.dong@intel.com> Signed-off-by: Avi Kivity <avi@qumranet.com>
Avi Kivity [Mon, 30 Apr 2007 11:47:02 +0000 (14:47 +0300)]
KVM: Assume that writes smaller than 4 bytes are to non-pagetable pages
This allows us to remove write protection earlier than otherwise. Should
some mad OS choose to use byte writes to update pagetables, it will suffer
a performance hit, but still work correctly.
KVM: VMX: Enable io bitmaps to avoid IO port 0x80 VMEXITs
This patch enables IO bitmaps control on vmx and unmask the 0x80 port to
avoid VMEXITs caused by accessing port 0x80. 0x80 is used as delays (see
include/asm/io.h), and handling VMEXITs on its access is unnecessary but
slows things down. This patch improves kernel build test at around
3%~5%.
Because every VM uses the same io bitmap, it is shared between
all VMs rather than a per-VM data structure.
Signed-off-by: Qing He <qing.he@intel.com> Signed-off-by: Avi Kivity <avi@qumranet.com>
* git://git.infradead.org/battery-2.6:
git-battery vs git-acpi
Power supply class and drivers: remove non obligatory return statements
pda_power: clean up irq, timer
MAINTAINERS: Add maintainers for power supply subsystem and drivers
Fixed up trivial conflict in drivers/w1/slaves/w1_ds2760.c manually
* master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6: (166 commits)
[SCSI] ibmvscsi: convert to use the data buffer accessors
[SCSI] dc395x: convert to use the data buffer accessors
[SCSI] ncr53c8xx: convert to use the data buffer accessors
[SCSI] sym53c8xx: convert to use the data buffer accessors
[SCSI] ppa: coding police and printk levels
[SCSI] aic7xxx_old: remove redundant GFP_ATOMIC from kmalloc
[SCSI] i2o: remove redundant GFP_ATOMIC from kmalloc from device.c
[SCSI] remove the dead CYBERSTORMIII_SCSI option
[SCSI] don't build scsi_dma_{map,unmap} for !HAS_DMA
[SCSI] Clean up scsi_add_lun a bit
[SCSI] 53c700: Remove printk, which triggers because of low scsi clock on SNI RMs
[SCSI] sni_53c710: Cleanup
[SCSI] qla4xxx: Fix underrun/overrun conditions
[SCSI] megaraid_mbox: use mutex instead of semaphore
[SCSI] aacraid: add 51245, 51645 and 52245 adapters to documentation.
[SCSI] qla2xxx: update version to 8.02.00-k1.
[SCSI] qla2xxx: add support for NPIV
[SCSI] stex: use resid for xfer len information
[SCSI] Add Brownie 1200U3P to blacklist
[SCSI] scsi.c: convert to use the data buffer accessors
...
Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6: (53 commits)
[TCP]: Verify the presence of RETRANS bit when leaving FRTO
[IPV6]: Call inet6addr_chain notifiers on link down
[NET_SCHED]: Kill CONFIG_NET_CLS_POLICE
[NET_SCHED]: act_api: qdisc internal reclassify support
[NET_SCHED]: sch_dsmark: act_api support
[NET_SCHED]: sch_atm: act_api support
[NET_SCHED]: sch_atm: Lindent
[IPV6]: MSG_ERRQUEUE messages do not pass to connected raw sockets
[IPV4]: Cleanup call to __neigh_lookup()
[NET_SCHED]: Revert "avoid transmit softirq on watchdog wakeup" optimization
[NETFILTER]: nf_conntrack: UDPLITE support
[NETFILTER]: nf_conntrack: mark protocols __read_mostly
[NETFILTER]: x_tables: add connlimit match
[NETFILTER]: Lower *tables printk severity
[NETFILTER]: nf_conntrack: Don't track locally generated special ICMP error
[NETFILTER]: nf_conntrack: Introduces nf_ct_get_tuplepr and uses it
[NETFILTER]: nf_conntrack: make l3proto->prepare() generic and renames it
[NETFILTER]: nf_conntrack: Increment error count on parsing IPv4 header
[NET]: Add ethtool support for NETIF_F_IPV6_CSUM devices.
[AF_IUCV]: Add lock when updating accept_q
...
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
9p: fix a race condition bug in umount which caused a segfault
9p: re-enable mount time debug option
9p: cache meta-data when cache=loose
net/9p: set error to EREMOTEIO if trans->write returns zero
net/9p: change net/9p module name to 9pnet
9p: Reorganization of 9p file system code
Merge branch 'for-linus' of git://oss.sgi.com:8090/xfs/xfs-2.6
* 'for-linus' of git://oss.sgi.com:8090/xfs/xfs-2.6: (37 commits)
[XFS] Fix lockdep annotations for xfs_lock_inodes
[LIB]: export radix_tree_preload()
[XFS] Fix XFS_IOC_FSBULKSTAT{,_SINGLE} & XFS_IOC_FSINUMBERS in compat mode
[XFS] Compat ioctl handler for handle operations
[XFS] Compat ioctl handler for XFS_IOC_FSGEOMETRY_V1.
[XFS] Clean up function name handling in tracing code
[XFS] Quota inode has no parent.
[XFS] Concurrent Multi-File Data Streams
[XFS] Use uninitialized_var macro to stop warning about rtx
[XFS] XFS should not be looking at filp reference counts
[XFS] Use is_power_of_2 instead of open coding checks
[XFS] Reduce shouting by removing unnecessary macros from dir2 code.
[XFS] Simplify XFS min/max macros.
[XFS] Kill off xfs_count_bits
[XFS] Cancel transactions on xfs_itruncate_start error.
[XFS] Use do_div() on 64 bit types.
[XFS] Fix remount,readonly path to flush everything correctly.
[XFS] Cleanup inode extent size hint extraction
[XFS] Prevent ENOSPC from aborting transactions that need to succeed
[XFS] Prevent deadlock when flushing inodes on unmount
...
Al Viro [Sun, 15 Jul 2007 20:37:16 +0000 (21:37 +0100)]
make i2c-acorn tristate
It depends on tristate I2C and it's trivial to make modular. The
current Kconfig allows I2C=m, I2C_ACORN=y, which doesn't work at
all; alternatives are dependency on I2C=y and making I2C_ACORN
itself a tristate. The latter is the right thing to do...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Acked-by: Russell King <rmk+kernel@arm.linux.org.uk>
Al Viro [Sun, 15 Jul 2007 20:01:32 +0000 (21:01 +0100)]
icside: devm_iounmap() needs linux/io.h
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Acked-by: Russell King <rmk+kernel@arm.linux.org.uk>
Jeff Garzik [Sat, 14 Jul 2007 23:12:04 +0000 (19:12 -0400)]
pda_power: clean up irq, timer
Clean up pda_power interrupt handling:
Prior to this patch, the driver would pass information it needed
to the interrupt handler dev_id pointer, and then prompt forget it
ever did so, recreating that same information after a couple passes
through the timer-based state machine.
This patch removes the redundant checks by passing the
pda_power_supply[] pointer through the state machine. The current
code passed 'irq' through the state machine, as an index to recreate
the pointer, when we could more simply pass around the pointer itself.
This patch makes it easier to remove the 'irq' argument in the future,
in addition to cleaning up the driver today.
FUJITA Tomonori [Fri, 25 May 2007 15:32:58 +0000 (00:32 +0900)]
[SCSI] ibmvscsi: convert to use the data buffer accessors
- remove the unnecessary map_single path.
- convert to use the new accessors for the sg lists and the
parameters.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp> Cc: Santiago Leon <santil@us.ibm.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Alan Cox [Mon, 9 Jul 2007 19:00:10 +0000 (12:00 -0700)]
[SCSI] ppa: coding police and printk levels
Add printk levels
Clean up some oddities of formatting
Fix goto labels
Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
[SCSI] aic7xxx_old: remove redundant GFP_ATOMIC from kmalloc
drivers/scsi/aic7xxx_old.c:aic7xxx_slave_alloc() unnecessarily passes
GFP_ATOMIC (along with GFP_KERNEL) to kmalloc() from a context that is not
atomic. Remove the pointless GFP_ATOMIC.
Signed-off-by: Satyam Sharma <ssatyam@cse.iitk.ac.in> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Cc: Doug Ledford <dledford@redhat.com> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
[SCSI] i2o: remove redundant GFP_ATOMIC from kmalloc from device.c
drivers/message/i2o/device.c:i2o_parm_field_get() unnecessarily passes
GFP_ATOMIC (along with GFP_KERNEL) to kmalloc() from a context that is not
atomic. Remove the pointless GFP_ATOMIC.
Signed-off-by: Satyam Sharma <ssatyam@cse.iitk.ac.in> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Adrian Bunk [Mon, 9 Jul 2007 19:00:10 +0000 (12:00 -0700)]
[SCSI] remove the dead CYBERSTORMIII_SCSI option
Not converted to the 2.6 kconfig system and no code in the tree.
Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
[IPV6]: Call inet6addr_chain notifiers on link down
Currently if the link is brought down via ip link or ifconfig down,
the inet6addr_chain notifiers are not called even though all
the addresses are removed from the interface. This caused SCTP
to add duplicate addresses to it's list.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Patrick McHardy [Sun, 15 Jul 2007 07:03:05 +0000 (00:03 -0700)]
[NET_SCHED]: Kill CONFIG_NET_CLS_POLICE
The NET_CLS_ACT option is now a full replacement for NET_CLS_POLICE,
remove the old code. The config option will be kept around to select
the equivalent NET_CLS_ACT options for a short time to allow easier
upgrades.
Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Patrick McHardy [Sun, 15 Jul 2007 07:02:31 +0000 (00:02 -0700)]
[NET_SCHED]: act_api: qdisc internal reclassify support
The behaviour of NET_CLS_POLICE for TC_POLICE_RECLASSIFY was to return
it to the qdisc, which could handle it internally or ignore it. With
NET_CLS_ACT however, tc_classify starts over at the first classifier
and never returns it to the qdisc. This makes it impossible to support
qdisc-internal reclassification, which in turn makes it impossible to
remove the old NET_CLS_POLICE code without breaking compatibility since
we have two qdiscs (CBQ and ATM) that support this.
This patch adds a tc_classify_compat function that handles
reclassification the old way and changes CBQ and ATM to use it.
This again is of course not fully backwards compatible with the previous
NET_CLS_ACT behaviour. Unfortunately there is no way to fully maintain
compatibility *and* support qdisc internal reclassification with
NET_CLS_ACT, but this seems like the better choice over keeping the two
incompatible options around forever.
Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Patrick McHardy [Sun, 15 Jul 2007 07:01:49 +0000 (00:01 -0700)]
[NET_SCHED]: sch_atm: act_api support
Handle act_api classification results.
The ATM scheduler behaves slightly different than other schedulers
in that it only handles policer results for successful classifications,
this behaviour is retained for the act_api case.
Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
[IPV6]: MSG_ERRQUEUE messages do not pass to connected raw sockets
From: Dmitry Butskoy <dmitry@butskoy.name>
Taken from http://bugzilla.kernel.org/show_bug.cgi?id=8747
Problem Description:
It is related to the possibility to obtain MSG_ERRQUEUE messages from the udp
and raw sockets, both connected and unconnected.
There is a little typo in net/ipv6/icmp.c code, which prevents such messages
to be delivered to the errqueue of the correspond raw socket, when the socket
is CONNECTED. The typo is due to swap of local/remote addresses.
Consider __raw_v6_lookup() function from net/ipv6/raw.c. When a raw socket is
looked up usual way, it is something like:
sk = __raw_v6_lookup(sk, nexthdr, daddr, saddr, IP6CB(skb)->iif);
where "daddr" is a destination address of the incoming packet (IOW our local
address), "saddr" is a source address of the incoming packet (the remote end).
But when the raw socket is looked up for some icmp error report, in
net/ipv6/icmp.c:icmpv6_notify() , daddr/saddr are obtained from the echoed
fragment of the "bad" packet, i.e. "daddr" is the original destination
address of that packet, "saddr" is our local address. Hence, for
icmpv6_notify() must use "saddr, daddr" in its arguments, not "daddr, saddr"
...
Steps to reproduce:
Create some raw socket, connect it to an address, and cause some error
situation: f.e. set ttl=1 where the remote address is more than 1 hop to reach.
Set IPV6_RECVERR .
Then send something and wait for the error (f.e. poll() with POLLERR|POLLIN).
You should receive "time exceeded" icmp message (because of "ttl=1"), but the
socket do not receive it.
If you do not connect your raw socket, you will receive MSG_ERRQUEUE
successfully. (The reason is that for unconnected socket there are no actual
checks for local/remote addresses).
Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Jean Delvare [Sun, 15 Jul 2007 03:51:44 +0000 (20:51 -0700)]
[IPV4]: Cleanup call to __neigh_lookup()
Back in the times of Linux 2.2, negative values for the creat parameter
of __neigh_lookup() had a particular meaning, but no longer, so we
should pass 1 instead.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Patrick McHardy [Sun, 15 Jul 2007 03:49:26 +0000 (20:49 -0700)]
[NET_SCHED]: Revert "avoid transmit softirq on watchdog wakeup" optimization
As noticed by Ranko Zivojnovic <ranko@spidernet.net>, calling qdisc_run
from the timer handler can result in deadlock:
> CPU#0
>
> qdisc_watchdog() fires and gets dev->queue_lock
> qdisc_run()...qdisc_restart()...
> -> releases dev->queue_lock and enters dev_hard_start_xmit()
>
> CPU#1
>
> tc del qdisc dev ...
> qdisc_graft()...dev_graft_qdisc()...dev_deactivate()...
> -> grabs dev->queue_lock ...
>
> qdisc_reset()...{cbq,hfsc,htb,netem,tbf}_reset()...qdisc_watchdog_cancel()...
> -> hrtimer_cancel() - waiting for the qdisc_watchdog() to exit, while still
> holding dev->queue_lock
>
> CPU#0
>
> dev_hard_start_xmit() returns ...
> -> wants to get dev->queue_lock(!)
>
> DEADLOCK!
The entire optimization is a bit questionable IMO, it moves potentially
large parts of NET_TX_SOFTIRQ work to TIMER_SOFTIRQ/HRTIMER_SOFTIRQ,
which kind of defeats the separation of them.
Signed-off-by: Patrick McHardy <kaber@trash.net> Acked-by: Ranko Zivojnovic <ranko@spidernet.net> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / linux-beck.git / log