** Changed in: ubuntu-z-systems
Status: Triaged => In Progress
--
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1903682
Title:
NULL pointer dereference when configuring multi-function with devfn !=
0 before devfn == 0
Status in Ubuntu on IBM z Systems:
In Progress
Status in linux package in Ubuntu:
Triaged
Status in linux source package in Focal:
Fix Committed
Status in linux source package in Groovy:
Triaged
Status in linux source package in Hirsute:
Triaged
Bug description:
SRU Justification:
==================
[Impact]
* While handling multifunction devices in zPCI the UID of the PCI
function with function number 0 (that always exists according to the
PCI spec) is taken as domain number.
* Therefore if hot plugging functions with a function number larger
than 0 are used before function 0, these need to be held in standby
before creating the domain and bus.
* This has been tested during development of this feature using a
patched QEMU and in DPM, but unfortunately never in
classic/traditional HMC mode.
* On a classic/traditional mode machine with a multi-function device,
and hot plug ("Reassign I/O Path") of the FID of the second port of
the LPAR, any additional hotplug (and even just deconfiguring a PCI
device) will hang - and hotplug now makes the entire Linux instance
unresponsive.
* The reason for this is a NULL pointer dereference - inc case
configuring multi-function with devfn != 0 before devfn == 0.
* This issue was introduced with the topology-aware PCI enumeration
code.
[Fix]
* 0b2ca2c7d0c9e2731d01b6c862375d44a7e13923 0b2ca2c7d0c9 "s390/pci: fix
hot-plug of PCI function missing bus"
[Test Case]
* IBM Z or LinuxONE hardware, equipped with hot-pluggable, multi-
functional PCIe cards (like for example RoCE Express 2 adapters) in
classic/traditional mode.
* An Ubuntu OS running in LPAR, that comes with a kernel that includes
the topology-aware PCI enumeration code (like for example 20.04.1 w/o
further updates or 20.10 GA kernel).
* Now on a system that is in classic/traditional mode, hot plug
("Reassign I/O Path") a multi-function device, but using the FID of
the second port.
[Regression Potential]
* There is at least some regression risk, but I consider it as low,
because:
* Even is the modification is a single if statement (that spans two
lines) in 'zpci_event_availability' it could harm the zPCI event
management even more, in worst case it could break hot plug not only
for systems in classic/traditional mode, but also in DPM mode (and
making the system hang) or for all ports.
* In such a case no enabling / disabling of devices would be possible.
* But the fix is very simple and straight-forward, it checks
zdev->zbus->bus for being NULL and in such a case break the function -
means breaking instead of calling the PCI common code
pci_scan_single_device() with the NULL pointer.
* PCIe devices are usually more optional devices on s390x (compared to
CCW and OSA devices for network) and this affects the zPCI subsystem
only, which is unique to s390x.
[Other]
* The patch got upstream accepted with kernel v5.10-rc3, hence it will
land sooner or later in Hirsute.
* The patch has also been tagged for the upstream stable v5.8 series,
hence will land in Groovy (based on kernel teams regular 'Groovy
update: v5.8.x upstream stable release' LP bug).
* Hence requesting this Kernel SRU for Focal only, since Ubuntu
releases older than Focal do not have the topology-aware zPCI
enumeration code.
__________
Background:
When handling multifunction devices in zPCI we take the
UID of the PCI function with function number 0
(that always exists according to the PCI spec)
as domain number.
Therefore when hot plugging functions with function
number larger than 0 before function 0, we need
to hold these in standby before creating the
domain and bus.
This has been tested during feature development
using a patched QEMU and with DPM but never in Classic
Mode.
Reproduction:
This issue was introduced with the Topology aware PCI
Enumeration code so test with a Linux supporting
that feature. E.g. Upstream, Devel Driver etc.
On a Classic Mode machine with a multi-function device,
hot plug ("Reassign I/O Path") only the FID of the
second port to the LPAR.
Symptom:
After this any additional hotplug and even just
deconfiguring a PCI device will hang. A hotplug
makes the entire Linux instance unresponsive.
Analysis:
The problem occurs in Classic Mode but not with
previous testing as the LPAR hypervisor does
hot plug/Reassign I/O Path as a two step process:
1. zPCI event with PEC 0x0302 to plug the zPCI function in Standby
2. zPCI event with PEC 0x0301 to configure the zPCI function
For the first event we create the zdev in clp_add_pci_device()
in Standby which is all fine so far.
The problem then occurs in step 2 as we then find
the existing zdev and try to configure it.
This however does not work as the PCI bus
is not yet created (as we still don't know the UID of
function 0 that will become its domain).
The bus pointer zdev->zbus->bus pointer is thus still
NULL but will be accessed by common code which
inevitably results in disaster including
the above mentioned hang and (possibly) the below
RCU stall:
[ 689.724703] rcu: INFO: rcu_sched self-detected stall on CPU
[ 689.724712] rcu: 16-....: (42004 ticks this GP)
idle=6ee/1/0x4000000000000002 softirq=1234/1234 fqs=14001
[ 689.724742] (t=42006 jiffies g=89 q=3770)
[ 689.724743] Task dump for CPU 16:
[ 689.724745] task:kmcheck state:R running task stack: 0
pid: 205 ppid: 2 flags:0x00000004
[ 689.724747] Call Trace:
[ 689.724757] [<0000000ccde0b5c4>] show_stack+0x8c/0xd8
[ 689.724762] [<0000000ccd0dabc4>] sched_show_task.part.0+0xe4/0x110
[ 689.724764] [<0000000ccde0ea5e>] rcu_dump_cpu_stacks+0xde/0x120
[ 689.724767] [<0000000ccd1465c6>] print_cpu_stall+0x266/0x330
[ 689.724768] [<0000000ccd14a428>] rcu_sched_clock_irq+0x618/0x670
[ 689.724771] [<0000000ccd15cd7a>] update_process_times+0xba/0xf0
[ 689.724775] [<0000000ccd1766fa>] tick_sched_timer+0x9a/0x220
[ 689.724777] [<0000000ccd15d962>] __hrtimer_run_queues+0x182/0x3a0
[ 689.724779] [<0000000ccd1602f8>] hrtimer_interrupt+0x138/0x450
[ 689.724782] [<0000000ccd0451c0>] do_IRQ+0x90/0xa0
[ 689.724784] [<0000000ccde2be96>] ext_int_handler+0x17e/0x184
[ 689.724790] [<0000000ccd9f373e>] pci_get_slot+0x5e/0xa0
[ 689.724794] [<0000000ccd9dc182>] pci_scan_single_device+0x32/0x2a0
[ 689.724797] [<0000000ccd0868f2>] __zpci_event_availability+0x192/0x360
[ 689.724800] [<0000000ccdd40c16>] chsc_process_crw+0x2e6/0x300
[ 689.724802] [<0000000ccdd4b088>] crw_collect_info+0x2b8/0x320
[ 689.724804] [<0000000ccd0caf3a>] kthread+0x14a/0x170
[ 689.724805] [<0000000ccde2b814>] ret_from_fork+0x24/0x2c
The fix is very simple, we check zdev->zbus->bus
for being NULL and in that case bail from the
case 0x0301 before calling the PCI common code
pci_scan_single_device() with the NULL pointer.
The only subtlety is that we still need to
do the zpci_enable_device() because the
code in arch/s390/pci/pci_bus.c assumes
that it can immediately do a scan of
all devfn != 0 PCI functions once
PCI function 0 is found.
It thereby mimics what happens
when we only find the FID for a function with
devfn != 0 in the CLP List PCI Functions.
This is implemented in the following upstream
commit:
0b2ca2c7d0c9e2731d01b6c862375d44a7e13923 s390/pci: fix hot-plug of PCI
function missing bus
It is included in v5.10-rc3 and has been tagged for
stable > v5.8 i.e. all upstream versions with
the PCI enumeration changes.
Also it carries the appropriate Fixes tag.
I have verified that it cherry-picks cleanly
on current focal master-next and expect
it to cleanly cherry-pick on newer Ubuntu
Kernels too.
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu-z-systems/+bug/1903682/+subscriptions
--
Mailing list: https://launchpad.net/~kernel-packages
Post to : kernel-packages@lists.launchpad.net
Unsubscribe : https://launchpad.net/~kernel-packages
More help : https://help.launchpad.net/ListHelp
