This patch fixes races caused by timed out ACPI IPMI transfers.
This patch uses timeout mechanism provided by ipmi_si to avoid the race
that the msg_done flag is set but without any protection, its content can
be invalid. Thanks for the suggestion of Corey Minyard.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Reviewed-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch fixes races caused by unprotected ACPI IPMI transfers.
We can see that the following crashes may occur:
1. There is no tx_msg_lock held for iterating tx_msg_list in
ipmi_flush_tx_msg() while it may be unlinked on failure in
parallel in acpi_ipmi_space_handler() under tx_msg_lock.
2. There is no lock held for freeing tx_msg in acpi_ipmi_space_handler()
while it may be accessed in parallel in ipmi_flush_tx_msg() and
ipmi_msg_handler().
This patch enhances tx_msg_lock to protect all tx_msg accesses to solve
this issue. Then tx_msg_lock is always held around complete() and tx_msg
accesses.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Reviewed-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch enhances sanity checks on message size to avoid potential buffer
overflow.
The kernel IPMI message size is IPMI_MAX_MSG_LENGTH(272 bytes) while the
ACPI specification defined IPMI message size is 64 bytes. The difference
is not handled by the original codes. This may cause crash in the response
handling codes.
This patch closes this gap and also combines rx_data/tx_data to use single
data/len pair since they need not be seperate.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Reviewed-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch fixes the issues indicated by the test results that
ipmi_msg_handler() is invoked in atomic context.
BUG: scheduling while atomic: kipmi0/18933/0x10000100
Modules linked in: ipmi_si acpi_ipmi ...
CPU: 3 PID: 18933 Comm: kipmi0 Tainted: G AW 3.10.0-rc7+ #2
Hardware name: QCI QSSC-S4R/QSSC-S4R, BIOS QSSC-S4R.QCI.01.00.0027.070120100606 07/01/2010
ffff8838245eea00 ffff88103fc63c98 ffffffff814c4a1e ffff88103fc63ca8
ffffffff814bfbab ffff88103fc63d28 ffffffff814c73e0 ffff88103933cbd4
0000000000000096 ffff88103fc63ce8 ffff88102f618000 ffff881035c01fd8
Call Trace:
<IRQ> [<ffffffff814c4a1e>] dump_stack+0x19/0x1b
[<ffffffff814bfbab>] __schedule_bug+0x46/0x54
[<ffffffff814c73e0>] __schedule+0x83/0x59c
[<ffffffff81058853>] __cond_resched+0x22/0x2d
[<ffffffff814c794b>] _cond_resched+0x14/0x1d
[<ffffffff814c6d82>] mutex_lock+0x11/0x32
[<ffffffff8101e1e9>] ? __default_send_IPI_dest_field.constprop.0+0x53/0x58
[<ffffffffa09e3f9c>] ipmi_msg_handler+0x23/0x166 [ipmi_si]
[<ffffffff812bf6e4>] deliver_response+0x55/0x5a
[<ffffffff812c0fd4>] handle_new_recv_msgs+0xb67/0xc65
[<ffffffff81007ad1>] ? read_tsc+0x9/0x19
[<ffffffff814c8620>] ? _raw_spin_lock_irq+0xa/0xc
[<ffffffffa09e1128>] ipmi_thread+0x5c/0x146 [ipmi_si]
...
Also Tony Camuso says:
We were getting occasional "Scheduling while atomic" call traces
during boot on some systems. Problem was first seen on a Cisco C210
but we were able to reproduce it on a Cisco c220m3. Setting
CONFIG_LOCKDEP and LOCKDEP_SUPPORT to 'y' exposed a lockdep around
tx_msg_lock in acpi_ipmi.c struct acpi_ipmi_device.
=================================
[ INFO: inconsistent lock state ]
2.6.32-415.el6.x86_64-debug-splck #1
---------------------------------
inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage.
ksoftirqd/3/17 [HC0[0]:SC1[1]:HE1:SE0] takes:
(&ipmi_device->tx_msg_lock){+.?...}, at: [<ffffffff81337a27>] ipmi_msg_handler+0x71/0x126
{SOFTIRQ-ON-W} state was registered at:
[<ffffffff810ba11c>] __lock_acquire+0x63c/0x1570
[<ffffffff810bb0f4>] lock_acquire+0xa4/0x120
[<ffffffff815581cc>] __mutex_lock_common+0x4c/0x400
[<ffffffff815586ea>] mutex_lock_nested+0x4a/0x60
[<ffffffff8133789d>] acpi_ipmi_space_handler+0x11b/0x234
[<ffffffff81321c62>] acpi_ev_address_space_dispatch+0x170/0x1be
The fix implemented by this change has been tested by Tony:
Tested the patch in a boot loop with lockdep debug enabled and never
saw the problem in over 400 reboots.
Reported-and-tested-by: Tony Camuso <tcamuso@redhat.com>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Reviewed-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
ACPI 4.0 spec adds the ACPI IPMI opregion, which means that the ACPI AML
code can also communicate with the BMC controller. This is to install
the ACPI IPMI opregion and enable the ACPI to access the BMC controller
through the IPMI message.
It will create IPMI user interface for every IPMI device detected
in ACPI namespace and install the corresponding IPMI opregion space handler.
Then it can enable ACPI to access the BMC controller through the IPMI
message.
The following describes how to process the IPMI request in IPMI space handler:
1. format the IPMI message based on the request in AML code.
IPMI system address. Now the address type is SYSTEM_INTERFACE_ADDR_TYPE
IPMI net function & command
IPMI message payload
2. send the IPMI message by using the function of ipmi_request_settime
3. wait for the completion of IPMI message. It can be done in different
routes: One is in handled in IPMI user recv callback function. Another is
handled in timeout function.
4. format the IPMI response and return it to ACPI AML code.
At the same time it also addes the module dependency. The ACPI IPMI opregion
will depend on the IPMI subsystem.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
cc: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Len Brown <len.brown@intel.com>
Lv Zheng