[ Upstream commit 95767ed78a ]
The resend_msg() function cannot fail, but there was error handling
around using it. Rework the handling of the error, and fix the out of
retries debug reporting that was wrong around this, too.
Cc: stable@vger.kernel.org
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Stable-dep-of: 00bb7e763e ("ipmi:ssif: Add a timer between request retries")
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 39721d62bb ]
The spec states that the minimum message retry time is 60ms, but it was
set to 20ms. Correct it.
Reported by: Tony Camuso <tcamuso@redhat.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Stable-dep-of: 00bb7e763e ("ipmi:ssif: Add a timer between request retries")
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit db05ddf7f3 upstream.
You will get two decrements when the messages on a panic are sent, not
one, since commit 2033f68589 ("ipmi: Free receive messages when in an
oops") was added, but the watchdog code had a bug where it didn't set
the value properly.
Reported-by: Anton Lundin <glance@acc.umu.se>
Cc: <Stable@vger.kernel.org> # v5.4+
Fixes: 2033f68589 ("ipmi: Free receive messages when in an oops")
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a92ce570c8 upstream.
The intf_free() function frees the "intf" pointer so we cannot
dereference it again on the next line.
Fixes: cbb79863fc ("ipmi: Don't allow device module unload when in use")
Signed-off-by: Dan Carpenter <error27@gmail.com>
Message-Id: <Y3M8xa1drZv4CToE@kili>
Cc: <stable@vger.kernel.org> # 5.5+
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit f6f1234d98 upstream.
When fixing the problem mentioned in PATCH1, we also found
the following problem:
If the IPMI is disconnected and in the sending process, the
uninstallation driver will be stuck for a long time.
The main problem is that uninstalling the driver waits for curr_msg to
be sent or HOSED. After stopping tasklet, the only place to trigger the
timeout mechanism is the circular poll in shutdown_smi.
The poll function delays 10us and calls smi_event_handler(smi_info,10).
Smi_event_handler deducts 10us from kcs->ibf_timeout.
But the poll func is followed by schedule_timeout_uninterruptible(1).
The time consumed here is not counted in kcs->ibf_timeout.
So when 10us is deducted from kcs->ibf_timeout, at least 1 jiffies has
actually passed. The waiting time has increased by more than a
hundredfold.
Now instead of calling poll(). call smi_event_handler() directly and
calculate the elapsed time.
For verification, you can directly use ebpf to check the kcs->
ibf_timeout for each call to kcs_event() when IPMI is disconnected.
Decrement at normal rate before unloading. The decrement rate becomes
very slow after unloading.
$ bpftrace -e 'kprobe:kcs_event {printf("kcs->ibftimeout : %d\n",
*(arg0+584));}'
Signed-off-by: Zhang Yuchen <zhangyuchen.lcr@bytedance.com>
Message-Id: <20221007092617.87597-3-zhangyuchen.lcr@bytedance.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 36992eb6b9 ]
After the IPMI disconnect problem, the memory kept rising and we tried
to unload the driver to free the memory. However, only part of the
free memory is recovered after the driver is uninstalled. Using
ebpf to hook free functions, we find that neither ipmi_user nor
ipmi_smi_msg is free, only ipmi_recv_msg is free.
We find that the deliver_smi_err_response call in clean_smi_msgs does
the destroy processing on each message from the xmit_msg queue without
checking the return value and free ipmi_smi_msg.
deliver_smi_err_response is called only at this location. Adding the
free handling has no effect.
To verify, try using ebpf to trace the free function.
$ bpftrace -e 'kretprobe:ipmi_alloc_recv_msg {printf("alloc rcv
%p\n",retval);} kprobe:free_recv_msg {printf("free recv %p\n",
arg0)} kretprobe:ipmi_alloc_smi_msg {printf("alloc smi %p\n",
retval);} kprobe:free_smi_msg {printf("free smi %p\n",arg0)}'
Signed-off-by: Zhang Yuchen <zhangyuchen.lcr@bytedance.com>
Message-Id: <20221007092617.87597-4-zhangyuchen.lcr@bytedance.com>
[Fixed the comment above handle_one_recv_msg().]
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 2ebaf18a0b ]
The was it was wouldn't work in some situations, simplify it. What was
there was unnecessary complexity.
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 7602b957e2 ]
Even though it's not possible to get into the SSIF_GETTING_MESSAGES and
SSIF_GETTING_EVENTS states without a valid message in the msg field,
it's probably best to be defensive here and check and print a log, since
that means something else went wrong.
Also add a default clause to that switch statement to release the lock
and print a log, in case the state variable gets messed up somehow.
Reported-by: Haowen Bai <baihaowen@meizu.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 34f35f8f14 upstream.
During probe ssif_info->client is dereferenced in error path. However,
it is set when some of the error checking has already been done. This
causes following kernel crash if an error path is taken:
[ 30.645593][ T674] ipmi_ssif 0-000e: ipmi_ssif: Not probing, Interface already present
[ 30.657616][ T674] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000088
...
[ 30.657723][ T674] pc : __dev_printk+0x28/0xa0
[ 30.657732][ T674] lr : _dev_err+0x7c/0xa0
...
[ 30.657772][ T674] Call trace:
[ 30.657775][ T674] __dev_printk+0x28/0xa0
[ 30.657778][ T674] _dev_err+0x7c/0xa0
[ 30.657781][ T674] ssif_probe+0x548/0x900 [ipmi_ssif 62ce4b08badc1458fd896206d9ef69a3c31f3d3e]
[ 30.657791][ T674] i2c_device_probe+0x37c/0x3c0
...
Initialize ssif_info->client before any error path can be taken. Clear
i2c_client data in the error path to prevent the dangling pointer from
leaking.
Fixes: c4436c9149 ("ipmi_ssif: avoid registering duplicate ssif interface")
Cc: stable@vger.kernel.org # 5.4.x
Suggested-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Mian Yousaf Kaukab <ykaukab@suse.de>
Message-Id: <20211208093239.4432-1-ykaukab@suse.de>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1d49eb91e8 upstream.
Currently when removing an ipmi_user the removal is deferred as a work on
the system's workqueue. Although this guarantees the free operation will
occur in non atomic context, it can race with the ipmi_msghandler module
removal (see [1]) . In case a remove_user work is scheduled for removal
and shortly after ipmi_msghandler module is removed we can end up in a
situation where the module is removed fist and when the work is executed
the system crashes with :
BUG: unable to handle page fault for address: ffffffffc05c3450
PF: supervisor instruction fetch in kernel mode
PF: error_code(0x0010) - not-present page
because the pages of the module are gone. In cleanup_ipmi() there is no
easy way to detect if there are any pending works to flush them before
removing the module. This patch creates a separate workqueue and schedules
the remove_work works on it. When removing the module the workqueue is
drained when destroyed to avoid the race.
[1] https://bugs.launchpad.net/bugs/1950666
Cc: stable@vger.kernel.org # 5.1
Fixes: 3b9a907223 (ipmi: fix sleep-in-atomic in free_user at cleanup SRCU user->release_barrier)
Signed-off-by: Ioanna Alifieraki <ioanna-maria.alifieraki@canonical.com>
Message-Id: <20211115131645.25116-1-ioanna-maria.alifieraki@canonical.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit b36eb5e7b7 ]
Don't do kfree or other risky things when oops_in_progress is set.
It's easy enough to avoid doing them
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 2253042d86 upstream.
When an IPMI watchdog timer is being stopped in ipmi_close() or
ipmi_ioctl(WDIOS_DISABLECARD), the current watchdog action is updated to
WDOG_TIMEOUT_NONE and _ipmi_set_timeout(IPMI_SET_TIMEOUT_NO_HB) is called
to install this action. The latter function ends up invoking
__ipmi_set_timeout() which makes the actual 'Set Watchdog Timer' IPMI
request.
For IPMI 1.0, this operation results in fully stopping the watchdog timer.
For IPMI >= 1.5, function __ipmi_set_timeout() always specifies the "don't
stop" flag in the prepared 'Set Watchdog Timer' IPMI request. This causes
that the watchdog timer has its action correctly updated to 'none' but the
timer continues to run. A problem is that IPMI firmware can then still log
an expiration event when the configured timeout is reached, which is
unexpected because the watchdog timer was requested to be stopped.
The patch fixes this problem by not setting the "don't stop" flag in
__ipmi_set_timeout() when the current action is WDOG_TIMEOUT_NONE which
results in stopping the watchdog timer. This makes the behaviour for
IPMI >= 1.5 consistent with IPMI 1.0. It also matches the logic in
__ipmi_heartbeat() which does not allow to reset the watchdog if the
current action is WDOG_TIMEOUT_NONE as that would start the timer.
Signed-off-by: Petr Pavlu <petr.pavlu@suse.com>
Message-Id: <10a41bdc-9c99-089c-8d89-fa98ce5ea080@suse.com>
Cc: stable@vger.kernel.org
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Pull IPMI updates from Corey Minyard:
"Some minor bug fixes, return values, cleanups of prints, conversion of
tasklets to the new API.
The biggest change is retrying the initial information fetch from the
management controller. If that fails, the iterface is not operational,
and one group was having trouble with the management controller not
being ready when the OS started up. So a retry was added"
* tag 'for-linus-5.10-1' of git://github.com/cminyard/linux-ipmi:
ipmi_si: Fix wrong return value in try_smi_init()
ipmi: msghandler: Fix a signedness bug
ipmi: add retry in try_get_dev_id()
ipmi: Clean up some printks
ipmi:msghandler: retry to get device id on an error
ipmi:sm: Print current state when the state is invalid
ipmi: Reset response handler when failing to send the command
ipmi: add a newline when printing parameter 'panic_op' by sysfs
char: ipmi: convert tasklets to use new tasklet_setup() API
The type for the completion codes should be unsigned char instead of
char. If it is declared as a normal char then the conditions in
__get_device_id() are impossible because the IPMI_DEVICE_IN_FW_UPDATE_ERR
error codes are higher than 127.
drivers/char/ipmi/ipmi_msghandler.c:2449 __get_device_id()
warn: impossible condition '(bmc->cc == 209) => ((-128)-127 == 209)'
Fixes: f8910ffa81 ("ipmi:msghandler: retry to get device id on an error")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Message-Id: <20200918142756.GB909725@mwanda>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Use a retry machanism to give the BMC more opportunities to correctly
respond when we receive specific completion codes.
This is similar to what is done in __get_device_id().
Signed-off-by: Xianting Tian <tian.xianting@h3c.com>
Message-Id: <20200916062129.26129-1-tian.xianting@h3c.com>
[Moved GET_DEVICE_ID_MAX_RETRY to include/linux/ipmi.h, reworded some
text.]
Signed-off-by: Corey Minyard <cminyard@mvista.com>
We fail to get the BMCS's device id with low probability when loading
the ipmi driver and it causes BMC device registration failed. When this
issue occurs we got below kernel prints:
[Wed Sep 9 19:52:03 2020] ipmi_si IPI0001:00: IPMI message handler:
device id demangle failed: -22
[Wed Sep 9 19:52:03 2020] IPMI BT: using default values
[Wed Sep 9 19:52:03 2020] IPMI BT: req2rsp=5 secs retries=2
[Wed Sep 9 19:52:03 2020] ipmi_si IPI0001:00: Unable to get the
device id: -5
[Wed Sep 9 19:52:04 2020] ipmi_si IPI0001:00: Unable to register
device: error -5
When this issue happens, we want to manually unload the driver and try to
load it again, but it can't be unloaded by 'rmmod' as it is already 'in
use'.
We add a print in handle_one_recv_msg(), when this issue happens,
the msg we received is "Recv: 1c 01 d5", which means the data_len is 1,
data[0] is 0xd5 (completion code), which means "bmc cannot execute
command. Command, or request parameter(s), not supported in present
state". Debug code:
static int handle_one_recv_msg(struct ipmi_smi *intf,
struct ipmi_smi_msg *msg) {
printk("Recv: %*ph\n", msg->rsp_size, msg->rsp);
... ...
}
Then in ipmi_demangle_device_id(), it returned '-EINVAL' as 'data_len < 7'
and 'data[0] != 0'.
We created this patch to retry the get device id when this error
happens. We reproduced this issue again and the retry succeed on the
first retry, we finally got the correct msg and then all is ok:
Recv: 1c 01 00 01 81 05 84 02 af db 07 00 01 00 b9 00 10 00
So use a retry machanism in this patch to give bmc more opportunity to
correctly response kernel when we received specific completion codes.
Signed-off-by: Xianting Tian <tian.xianting@h3c.com>
Message-Id: <20200915071817.4484-1-tian.xianting@h3c.com>
[Cleaned up the verbage a bit in the header and prints.]
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Print current state before returning IPMI_NOT_IN_MY_STATE_ERR so we can
know where this issue is coming from and possibly fix the state machine.
Signed-off-by: Xianting Tian <tian.xianting@h3c.com>
Message-Id: <20200915074441.4090-1-tian.xianting@h3c.com>
[Converted printk() to pr_xxx().]
Signed-off-by: Corey Minyard <cminyard@mvista.com>
When I cat some ipmi_watchdog parameters by sysfs, it displays as
follows. It's better to add a newline for easy reading.
root@(none):/# cat /sys/module/ipmi_watchdog/parameters/action
resetroot@(none):/# cat /sys/module/ipmi_watchdog/parameters/preaction
pre_noneroot@(none):/# cat /sys/module/ipmi_watchdog/parameters/preop
preop_noneroot@(none):/#
Signed-off-by: Xiongfeng Wang <wangxiongfeng2@huawei.com>
Message-Id: <1595313309-43881-1-git-send-email-wangxiongfeng2@huawei.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Pull IPMI updates from Corey Minyard:
"A few small fixes for things, nothing earth shattering"
* tag 'for-linus-5.8-1' of git://github.com/cminyard/linux-ipmi:
ipmi:ssif: Remove dynamic platform device handing
Try to load acpi_ipmi when an SSIF ACPI IPMI interface is added
ipmi_si: Load acpi_ipmi when ACPI IPMI interface added
ipmi:bt-bmc: Fix error handling and status check
ipmi: Replace guid_copy() with import_guid() where it makes sense
ipmi: use vzalloc instead of kmalloc for user creation
ipmi:bt-bmc: Fix some format issue of the code
ipmi:bt-bmc: Avoid unnecessary check
Pull documentation updates from Jonathan Corbet:
"A fair amount of stuff this time around, dominated by yet another
massive set from Mauro toward the completion of the RST conversion. I
*really* hope we are getting close to the end of this. Meanwhile,
those patches reach pretty far afield to update document references
around the tree; there should be no actual code changes there. There
will be, alas, more of the usual trivial merge conflicts.
Beyond that we have more translations, improvements to the sphinx
scripting, a number of additions to the sysctl documentation, and lots
of fixes"
* tag 'docs-5.8' of git://git.lwn.net/linux: (130 commits)
Documentation: fixes to the maintainer-entry-profile template
zswap: docs/vm: Fix typo accept_threshold_percent in zswap.rst
tracing: Fix events.rst section numbering
docs: acpi: fix old http link and improve document format
docs: filesystems: add info about efivars content
Documentation: LSM: Correct the basic LSM description
mailmap: change email for Ricardo Ribalda
docs: sysctl/kernel: document unaligned controls
Documentation: admin-guide: update bug-hunting.rst
docs: sysctl/kernel: document ngroups_max
nvdimm: fixes to maintainter-entry-profile
Documentation/features: Correct RISC-V kprobes support entry
Documentation/features: Refresh the arch support status files
Revert "docs: sysctl/kernel: document ngroups_max"
docs: move locking-specific documents to locking/
docs: move digsig docs to the security book
docs: move the kref doc into the core-api book
docs: add IRQ documentation at the core-api book
docs: debugging-via-ohci1394.txt: add it to the core-api book
docs: fix references for ipmi.rst file
...
Platform devices can only come in through the DMI interface, and that
will get done before initialization is complete. Therefore there is no
reason to hande getting a device in new_ssif_client after
initialization.
Dynamic entries can still come in through the i2c interfaces, but that's
handled differently.
Signed-off-by: Corey Minyard <cminyard@mvista.com>
If the function platform_get_irq() failed, the negative value
returned will not be detected here. So fix error handling in
bt_bmc_config_irq(). And in the function bt_bmc_probe(),
when get irq failed, it will print error message. So use
platform_get_irq_optional() to simplify code. Finally in the
function bt_bmc_remove() should make the right status check
if get irq failed.
Signed-off-by: Shengju Zhang <zhangshengju@cmss.chinamobile.com>
Signed-off-by: Tang Bin <tangbin@cmss.chinamobile.com>
Message-Id: <20200505102906.17196-1-tangbin@cmss.chinamobile.com>
[Also set bt_bmc->irq to a negative value if devm_request_irq() fails.]
Signed-off-by: Corey Minyard <cminyard@mvista.com>
We met mulitple times of failure of staring bmc-watchdog,
due to the runtime memory allocation failure of order 4.
bmc-watchdog: page allocation failure: order:4, mode:0x40cc0(GFP_KERNEL|__GFP_COMP), nodemask=(null),cpuset=/,mems_allowed=0-1
CPU: 1 PID: 2571 Comm: bmc-watchdog Not tainted 5.5.0-00045-g7d6bb61d6188c #1
Hardware name: Intel Corporation S2600WFT/S2600WFT, BIOS SE5C620.86B.00.01.0015.110720180833 11/07/2018
Call Trace:
dump_stack+0x66/0x8b
warn_alloc+0xfe/0x160
__alloc_pages_slowpath+0xd3e/0xd80
__alloc_pages_nodemask+0x2f0/0x340
kmalloc_order+0x18/0x70
kmalloc_order_trace+0x1d/0xb0
ipmi_create_user+0x55/0x2c0 [ipmi_msghandler]
ipmi_open+0x72/0x110 [ipmi_devintf]
chrdev_open+0xcb/0x1e0
do_dentry_open+0x1ce/0x380
path_openat+0x305/0x14f0
do_filp_open+0x9b/0x110
do_sys_open+0x1bd/0x250
do_syscall_64+0x5b/0x1f0
entry_SYSCALL_64_after_hwframe+0x44/0xa9
Using vzalloc/vfree for creating ipmi_user heals the
problem
Thanks to Stephen Rothwell for finding the vmalloc.h
inclusion issue.
Signed-off-by: Feng Tang <feng.tang@intel.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
This needs to return the newly allocated struct but instead it returns
zero which leads to an immediate Oops in the caller.
Fixes: 09f5f68070 ("ipmi: kcs: aspeed: Implement v2 bindings")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Message-Id: <20200407122149.GA100026@mwanda>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Sparse reports a warning at ipmi_ssif_unlock_cond()
and ipmi_ssif_lock_cond()
warning: context imbalance in ipmi_ssif_lock_cond()
- wrong count at exit
warning: context imbalance in ipmi_ssif_unlock_cond()
- unexpected unlock
The root cause is the missing annotation at ipmi_ssif_unlock_cond()
and ipmi_ssif_lock_cond()
Add the missing __acquires(&ata_scsi_rbuf_lock)
Add the missing __releases(&ata_scsi_rbuf_lock)
Signed-off-by: Jules Irenge <jbi.octave@gmail.com>
Message-Id: <20200403160505.2832-6-jbi.octave@gmail.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
