12 Apr, 2014
1 commit
-
Pull 9p changes from Eric Van Hensbergen:
"A bunch of updates and cleanup within the transport layer,
particularly with a focus on RDMA"* tag 'for-linus-3.15' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
9pnet_rdma: check token type before int conversion
9pnet: trans_fd : allocate struct p9_trans_fd and struct p9_conn together.
9pnet: p9_client->conn field is unused. Remove it.
9P: Get rid of REQ_STATUS_FLSH
9pnet_rdma: add cancelled()
9pnet_rdma: update request status during send
9P: Add cancelled() to the transport functions.
net: Mark function as static in 9p/client.c
9P: Add memory barriers to protect request fields over cb/rpc threads handoff
26 Mar, 2014
8 commits
-
When parsing options, make sure we have found a proper token before
doing a numeric conversion.Without this check, the current code will end up following random
pointers that just happened to be on the stack when this function was
called, because match_token() will not touch the 'args' list unless a
valid token is found.Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
There is no point in allocating these structs separately.
Changing this makes the code a little simpler and saves a few bytes of
memory.Reported-by: Herve Vico
Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
This request state is mostly useless, and properly implementing it
for RDMA would require an extra lock to be taken in handle_recv()
and in rdma_cancel() to avoid this race:handle_recv() rdma_cancel()
. .
. if req->state == SENT
req->state = RCVD .
. req->state = FLSHSo just get rid of it.
Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
Take into account posted recv buffers that will never receive their
reply.The RDMA code posts a recv buffer for each request that it sends.
When a request is flushed, it is possible that this request will
never receive a reply, and that one recv buffer will stay unused on
the recv queue.It is then possible, if this scenario happens several times, to have the
recv queue full, and have the 9pnet_rmda module unable to send new requests.Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
This will be needed by the flush logic.
Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
And move transport-specific code out of net/9p/client.c
Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
Mark function as static in net/9p/client.c because it is not used
outside this file.This eliminates the following warning in net/9p/client.c:
net/9p/client.c:207:18: warning: no previous prototype for ‘p9_fcall_alloc’ [-Wmissing-prototypes]Signed-off-by: Rashika Kheria
Reviewed-by: Josh Triplett
Signed-off-by: Eric Van Hensbergen -
We need barriers to guarantee this pattern works as intended:
[w] req->rc, 1 [r] req->status, 1
wmb rmb
[w] req->status, 1 [r] req->rcWhere the wmb ensures that rc gets written before status,
and the rmb ensures that if you observe status == 1, rc is the new value.Signed-off-by: Dominique Martinet
Signed-off-by: Eric Van Hensbergen
11 Feb, 2014
1 commit
-
The 9p-virtio transport does zero copy on things larger than 1024 bytes
in size. It accomplishes this by returning the physical addresses of
pages to the virtio-pci device. At present, the translation is usually a
bit shift.That approach produces an invalid page address when we read/write to
vmalloc buffers, such as those used for Linux kernel modules. Any
attempt to load a Linux kernel module from 9p-virtio produces the
following stack.[] p9_virtio_zc_request+0x45e/0x510
[] p9_client_zc_rpc.constprop.16+0xfd/0x4f0
[] p9_client_read+0x15d/0x240
[] v9fs_fid_readn+0x50/0xa0
[] v9fs_file_readn+0x10/0x20
[] v9fs_file_read+0x37/0x70
[] vfs_read+0x9b/0x160
[] kernel_read+0x41/0x60
[] copy_module_from_fd.isra.34+0xfb/0x180Subsequently, QEMU will die printing:
qemu-system-x86_64: virtio: trying to map MMIO memory
This patch enables 9p-virtio to correctly handle this case. This not
only enables us to load Linux kernel modules off virtfs, but also
enables ZFS file-based vdevs on virtfs to be used without killing QEMU.Special thanks to both Avi Kivity and Alexander Graf for their
interpretation of QEMU backtraces. Without their guidence, tracking down
this bug would have taken much longer. Also, special thanks to Linus
Torvalds for his insightful explanation of why this should use
is_vmalloc_addr() instead of is_vmalloc_or_module_addr():https://lkml.org/lkml/2014/2/8/272
Signed-off-by: Richard Yao
Signed-off-by: David S. Miller
10 Feb, 2014
1 commit
-
Mark function as static in net/9p/client.c because it is not used
outside this file.This eliminates the following warning in net/9p/client.c:
net/9p/client.c:207:18: warning: no previous prototype for ‘p9_fcall_alloc’ [-Wmissing-prototypes]Signed-off-by: Rashika Kheria
Reviewed-by: Josh Triplett
Signed-off-by: David S. Miller
24 Nov, 2013
1 commit
-
A few releases back a patch made virtio the default transport, however
it was done in a way which side-stepped the mechanism put in place to
allow for this selection. This patch cleans that up while maintaining
virtio as the default transport.Signed-off-by: Eric Van Hensbergen
15 Nov, 2013
1 commit
-
Pull virtio updates from Rusty Russell:
"Nothing really exciting: some groundwork for changing virtio endian,
and some robustness fixes for broken virtio devices, plus minor
tweaks"* tag 'virtio-next-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux:
virtio_scsi: verify if queue is broken after virtqueue_get_buf()
x86, asmlinkage, lguest: Pass in globals into assembler statement
virtio: mmio: fix signature checking for BE guests
virtio_ring: adapt to notify() returning bool
virtio_net: verify if queue is broken after virtqueue_get_buf()
virtio_console: verify if queue is broken after virtqueue_get_buf()
virtio_blk: verify if queue is broken after virtqueue_get_buf()
virtio_ring: add new function virtqueue_is_broken()
virtio_test: verify if virtqueue_kick() succeeded
virtio_net: verify if virtqueue_kick() succeeded
virtio_ring: let virtqueue_{kick()/notify()} return a bool
virtio_ring: change host notification API
virtio_config: remove virtio_config_val
virtio: use size-based config accessors.
virtio_config: introduce size-based accessors.
virtio_ring: plug kmemleak false positive.
virtio: pm: use CONFIG_PM_SLEEP instead of CONFIG_PM
25 Oct, 2013
1 commit
-
Signed-off-by: Al Viro
17 Oct, 2013
1 commit
-
This lets the transport do endian conversion if necessary, and insulates
the drivers from the difference.Most drivers can use the simple helpers virtio_cread() and virtio_cwrite().
Signed-off-by: Rusty Russell
12 Sep, 2013
1 commit
-
Pull 9p updates from Eric Van Hensbergen:
"Minor 9p fixes and tweaks for 3.12 merge windowThe first fixes namespace issues which causes a kernel NULL pointer
dereference, the second fixes uevent handling to work better with
udev, and the third switches some code to use srlcpy instead of
strncpy in order to be safer.All changes have been baking in for-next for at least 2 weeks"
* tag 'for-linus-3.12-merge' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
fs/9p: avoid accessing utsname after namespace has been torn down
9p: send uevent after adding/removing mount_tag attribute
fs: 9p: use strlcpy instead of strncpy
26 Aug, 2013
2 commits
-
During trinity fuzzing in a kvmtool guest, I stumbled across the
following:Unable to handle kernel NULL pointer dereference at virtual address 00000004
PC is at v9fs_file_do_lock+0xc8/0x1a0
LR is at v9fs_file_do_lock+0x48/0x1a0
[] (v9fs_file_do_lock+0xc8/0x1a0) from [] (locks_remove_flock+0x8c/0x124)
[] (locks_remove_flock+0x8c/0x124) from [] (__fput+0x58/0x1e4)
[] (__fput+0x58/0x1e4) from [] (task_work_run+0xac/0xe8)
[] (task_work_run+0xac/0xe8) from [] (do_exit+0x6bc/0x8d8)
[] (do_exit+0x6bc/0x8d8) from [] (do_group_exit+0x3c/0xb0)
[] (do_group_exit+0x3c/0xb0) from [] (__wake_up_parent+0x0/0x18)I believe this is due to an attempt to access utsname()->nodename, after
exit_task_namespaces() has been called, leaving current->nsproxy->uts_ns
as NULL and causing the above dereference.A similar issue was fixed for lockd in 9a1b6bf818e7 ("LOCKD: Don't call
utsname()->nodename from nlmclnt_setlockargs"), so this patch attempts
something similar for 9pfs.Cc: Eric Van Hensbergen
Cc: Ron Minnich
Cc: Latchesar Ionkov
Cc: Trond Myklebust
Signed-off-by: Will Deacon
Signed-off-by: Eric Van Hensbergen -
This driver adds an attribute to the existing virtio device so a CHANGE
event is required in order udev rules to make use of it. The ADD event
happens before this driver is probed and unlike a more typical driver
like a block device there isn't a higher level device to watch for.Signed-off-by: Michael Marineau
Signed-off-by: Eric Van Hensbergen
31 Jul, 2013
1 commit
-
This patch reverts commit
80b45261a0b263536b043c5ccfc4ba4fc27c2acc
which was implementing a 'cancelled' functionality to notify that
a cancelled request will not be replied.This implementation was not used anywhere and therefore removed.
Signed-off-by: Andi Shyti
Signed-off-by: David S. Miller
25 Jul, 2013
1 commit
-
This patch gets rid of the following warning:
net/9p/trans_rdma.c:594:12: warning: ‘rdma_cancelled’ defined but not used [-Wunused-function]
static int rdma_cancelled(struct p9_client *client, struct p9_req_t *req)The rdma_cancelled function is not called anywhere in the kernel
Signed-off-by: Andi Shyti
Signed-off-by: David S. Miller
14 Jul, 2013
1 commit
-
Pull networking fixes from David Miller:
"Just a bunch of small fixes and tidy ups:1) Finish the "busy_poll" renames, from Eliezer Tamir.
2) Fix RCU stalls in IFB driver, from Ding Tianhong.
3) Linearize buffers properly in tun/macvtap zerocopy code.
4) Don't crash on rmmod in vxlan, from Pravin B Shelar.
5) Spinlock used before init in alx driver, from Maarten Lankhorst.
6) A sparse warning fix in bnx2x broke TSO checksums, fix from Dmitry
Kravkov.7) Dummy and ifb driver load failure paths can oops, fixes from Tan
Xiaojun and Ding Tianhong.8) Correct MTU calculations in IP tunnels, from Alexander Duyck.
9) Account all TCP retransmits in SNMP stats properly, from Yuchung
Cheng.10) atl1e and via-rhine do not handle DMA mapping failures properly,
from Neil Horman.11) Various equal-cost multipath route fixes in ipv6 from Hannes
Frederic Sowa"* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (36 commits)
ipv6: only static routes qualify for equal cost multipathing
via-rhine: fix dma mapping errors
atl1e: fix dma mapping warnings
tcp: account all retransmit failures
usb/net/r815x: fix cast to restricted __le32
usb/net/r8152: fix integer overflow in expression
net: access page->private by using page_private
net: strict_strtoul is obsolete, use kstrtoul instead
drivers/net/ieee802154: don't use devm_pinctrl_get_select_default() in probe
drivers/net/ethernet/cadence: don't use devm_pinctrl_get_select_default() in probe
drivers/net/can/c_can: don't use devm_pinctrl_get_select_default() in probe
net/usb: add relative mii functions for r815x
net/tipc: use %*phC to dump small buffers in hex form
qlcnic: Adding Maintainers.
gre: Fix MTU sizing check for gretap tunnels
pkt_sched: sch_qfq: remove forward declaration of qfq_update_agg_ts
pkt_sched: sch_qfq: improve efficiency of make_eligible
gso: Update tunnel segmentation to support Tx checksum offload
inet: fix spacing in assignment
ifb: fix oops when loading the ifb failed
...
12 Jul, 2013
2 commits
-
p9_release_pages() would attempt to dereference one value past the end of
pages[]. This would cause the following crashes:[ 6293.171817] BUG: unable to handle kernel paging request at ffff8807c96f3000
[ 6293.174146] IP: [] p9_release_pages+0x3b/0x60
[ 6293.176447] PGD 79c5067 PUD 82c1e3067 PMD 82c197067 PTE 80000007c96f3060
[ 6293.180060] Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
[ 6293.180060] Modules linked in:
[ 6293.180060] CPU: 62 PID: 174043 Comm: modprobe Tainted: G W 3.10.0-next-20130710-sasha #3954
[ 6293.180060] task: ffff8807b803b000 ti: ffff880787dde000 task.ti: ffff880787dde000
[ 6293.180060] RIP: 0010:[] [] p9_release_pages+0x3b/0x60
[ 6293.214316] RSP: 0000:ffff880787ddfc28 EFLAGS: 00010202
[ 6293.214316] RAX: 0000000000000001 RBX: ffff8807c96f2ff8 RCX: 0000000000000000
[ 6293.222017] RDX: ffff8807b803b000 RSI: 0000000000000001 RDI: ffffea001c7e3d40
[ 6293.222017] RBP: ffff880787ddfc48 R08: 0000000000000000 R09: 0000000000000000
[ 6293.222017] R10: 0000000000000001 R11: 0000000000000000 R12: 0000000000000001
[ 6293.222017] R13: 0000000000000001 R14: ffff8807cc50c070 R15: ffff8807cc50c070
[ 6293.222017] FS: 00007f572641d700(0000) GS:ffff8807f3600000(0000) knlGS:0000000000000000
[ 6293.256784] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[ 6293.256784] CR2: ffff8807c96f3000 CR3: 00000007c8e81000 CR4: 00000000000006e0
[ 6293.256784] Stack:
[ 6293.256784] ffff880787ddfcc8 ffff880787ddfcc8 0000000000000000 ffff880787ddfcc8
[ 6293.256784] ffff880787ddfd48 ffffffff84128be8 ffff880700000002 0000000000000001
[ 6293.256784] ffff8807b803b000 ffff880787ddfce0 0000100000000000 0000000000000000
[ 6293.256784] Call Trace:
[ 6293.256784] [] p9_virtio_zc_request+0x598/0x630
[ 6293.256784] [] ? wake_up_bit+0x40/0x40
[ 6293.256784] [] p9_client_zc_rpc+0x111/0x3a0
[ 6293.256784] [] ? sched_clock_cpu+0x108/0x120
[ 6293.256784] [] p9_client_read+0xe1/0x2c0
[ 6293.256784] [] v9fs_file_read+0x90/0xc0
[ 6293.256784] [] vfs_read+0xc3/0x130
[ 6293.256784] [] ? trace_hardirqs_on+0xd/0x10
[ 6293.256784] [] SyS_read+0x62/0xa0
[ 6293.256784] [] tracesys+0xdd/0xe2
[ 6293.256784] Code: 66 90 48 89 fb 41 89 f5 48 8b 3f 48 85 ff 74 29 85 f6 74 25 45 31 e4 66 0f 1f 84 00 00 00 00 00 e8 eb 14 12 fd 41 ff c4 49 63 c4 8b 3c c3 48 85 ff 74 05 45 39 e5 75 e7 48 83 c4 08 5b 41 5c
[ 6293.256784] RIP [] p9_release_pages+0x3b/0x60
[ 6293.256784] RSP
[ 6293.256784] CR2: ffff8807c96f3000
[ 6293.256784] ---[ end trace 50822ee72cd360fc ]---Signed-off-by: Sasha Levin
Signed-off-by: David S. Miller -
…inux/kernel/git/ericvh/v9fs
Pull second round of 9p patches from Eric Van Hensbergen:
"Several of these patches were rebased in order to correct style
issues. Only stylistic changes were made versus the patches which
were in linux-next for two weeks. The rebases have been in linux-next
for 3 days and have passed my regressions.The bulk of these are RDMA fixes and improvements. There's also some
additions on the extended attributes front to support some additional
namespaces and a new option for TCP to force allocation of mount
requests from a priviledged port"* tag 'for-linus-3.11-merge-window-part-2' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
fs/9p: Remove the unused variable "err" in v9fs_vfs_getattr()
9P: Add cancelled() to the transport functions.
9P/RDMA: count posted buffers without a pending request
9P/RDMA: Improve error handling in rdma_request
9P/RDMA: Do not free req->rc in error handling in rdma_request()
9P/RDMA: Use a semaphore to protect the RQ
9P/RDMA: Protect against duplicate replies
9P/RDMA: increase P9_RDMA_MAXSIZE to 1MB
9pnet: refactor struct p9_fcall alloc code
9P/RDMA: rdma_request() needs not allocate req->rc
9P: Fix fcall allocation for rdma
fs/9p: xattr: add trusted and security namespaces
net/9p: add privport option to 9p tcp transport
10 Jul, 2013
1 commit
-
…inux/kernel/git/ericvh/v9fs
Pull 9p update from Eric Van Hensbergen:
"Grab bag of little fixes and enhancements:
- optional security enhancements
- fix path coverage in MAINTAINERS
- switch to using most used protocol and transport as default
- clean up buffer dumps in trace codeHeld off on RDMA patches as they need to be cleaned up a bit, but will
try to get the cleaned, checked, and pushed by mid-week"* tag 'for-linus-3.11-merge-window-part-1' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
9p: Add rest of 9p files to MAINTAINERS entry
9p: trace: use %*ph to dump buffer
net/9p: Handle error in zero copy request correctly for 9p2000.u
net/9p: Use virtio transpart as the default transport
net/9p: Make 9P2000.L the default protocol for 9p file system
08 Jul, 2013
11 commits
-
RDMA needs to post a buffer for each incoming reply.
Hence it needs to keep count of these and needs to be
aware of whether a flushed request has received a reply
or not.This patch adds the cancelled() callback to the transport modules.
It is called when RFLUSH has been received and that the corresponding
request will never receive a reply.Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
In rdma_request():
If an error occurs between posting the recv and the send,
there will be a reply context posted without a pending
request.
Since there is no way to "un-post" it, we remember it and
skip post_recv() for the next request.Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
Most importantly:
- do not free the recv context (rpl_context) after a successful post_recv()
- but do free the send context (c) after a failed send.Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
rdma_request() should never be in charge of freeing rc.
When an error occurs:
* Either the rc buffer has been recv_post()'ed.
then kfree()'ing it certainly is a bad idea.
* Or is has not, and in that case req->rc still points to it,
hence it needs not be freed.Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
The current code keeps track of the number of buffers posted in the RQ,
and will prevent it from overflowing. But it does so by simply dropping
post requests (And leaking memory in the process).
When this happens there will actually be too few buffers posted, and
soon the 9P server will complain about 'RNR retry counter exceeded'
errors.Instead, use a semaphore, and block until the RQ is ready for another
buffer to be posted.Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
A well-behaved server would not send twice the reply to a request.
But if it ever happens...
This additional check prevents the kernel from leaking memory
and possibly more nasty consequences in that unlikely event.Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
The current value is too low to get good performance.
Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
p9_tag_alloc() takes care of that.
Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
The current code assumes that when a request in the request array
does have a tc, it also has a rc.This is normally true, but not always : when using RDMA, req->rc
will temporarily be set to NULL after the request has been sent.
That is usually OK though, as when the reply arrives, req->rc will be
reassigned to a sane value before the request is recycled.But there is a catch : if the request is flushed, the reply will never
arrive, and req->rc will be NULL, but not req->tc.This patch fixes p9_tag_alloc to take this into account.
Signed-off-by: Simon Derr
Signed-off-by: Eric Van Hensbergen -
If the privport option is specified, the tcp transport binds local
address to a reserved port before connecting to the 9p server.In some cases when 9P AUTH cannot be implemented, this is better than
nothing.Signed-off-by: Jim Garlick
Signed-off-by: Eric Van Hensbergen
11 Jun, 2013
1 commit
-
Pull net/9p bug fix from Eric Van Hensbergen:
"zero copy error fix"* tag '9p-3.10-bug-fix-1' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
net/9p: Handle error in zero copy request correctly for 9p2000.u
29 May, 2013
1 commit
-
For zero copy request, error will be encoded in the user space buffer.
So copy the error code correctly using copy_from_user. Here we use the
extra bytes we allocate for zero copy request. If total error details
are more than P9_ZC_HDR_SZ - 7 bytes, we return -EFAULT. The patch also
avoid a memory allocation in the error path.Signed-off-by: Aneesh Kumar K.V
Signed-off-by: Eric Van Hensbergen
28 May, 2013
3 commits
-
For zero copy request, error will be encoded in the user space buffer.
So copy the error code correctly using copy_from_user. Here we use the
extra bytes we allocate for zero copy request. If total error details
are more than P9_ZC_HDR_SZ - 7 bytes, we return -EFAULT. The patch also
avoid a memory allocation in the error path.Signed-off-by: Aneesh Kumar K.V
Signed-off-by: Eric Van Hensbergen -
Make the default 9p experience better by defaulting to virtio transport if present.
These days most of the users are using 9p in a virtualized setupSigned-off-by: Aneesh Kumar K.V
Signed-off-by: Eric Van Hensbergen -
If we dont' specify a protocol version default to 9P2000.L. 9P2000.L
have better support for posix semantic and is where all the recent development
is happening.Signed-off-by: Aneesh Kumar K.V
Signed-off-by: Eric Van Hensbergen