Hi Peter,
On 8/2/24 11:49 AM, Peter Newman wrote:
On Fri, Aug 2, 2024 at 9:14 AM Reinette Chatre
<reinette.chatre@xxxxxxxxx> wrote:
On 8/1/24 3:45 PM, Peter Newman wrote:
On Thu, Aug 1, 2024 at 2:50 PM Reinette Chatre
<reinette.chatre@xxxxxxxxx> wrote:
On 7/17/24 10:19 AM, Moger, Babu wrote:
On 7/12/24 17:03, Reinette Chatre wrote:
On 7/3/24 2:48 PM, Babu Moger wrote:
# Examples
a. Check if ABMC support is available
#mount -t resctrl resctrl /sys/fs/resctrl/
#cat /sys/fs/resctrl/info/L3_MON/mbm_mode
[abmc]
legacy
Linux kernel detected ABMC feature and it is enabled.
How about renaming "abmc" to "mbm_cntrs"? This will match the num_mbm_cntrs
info file and be the final step to make this generic so that another
architecture
can more easily support assignining hardware counters without needing to call
the feature AMD's "abmc".
I think we aleady settled this with "mbm_cntr_assignable".
For soft-RMID" it will be mbm_sw_assignable.
Maybe getting a bit long but how about "mbm_cntr_sw_assignable" to match
with the term "mbm_cntr" in accompanying "num_mbm_cntrs"?
My users are pushing for a consistent interface regardless of whether
counter assignment is implemented in hardware or software, so I would
like to avoid exposing implementation differences in the interface
where possible.
This seems a reasonable ask but can we be confident that if hardware
supports assignable counters then there will never be a reason to use
software assignable counters? (This needs to also consider how/if Arm
may use this feature.)
I am of course not familiar with details of the software implementation
- could there be benefits to using it even if hardware counters are
supported?
I can't see any situation where the user would want to choose software
over hardware counters. The number of groups which can be monitored by
software assignable counters will always be less than with hardware,
due to the need for consuming one RMID (and the counters automatically
allocated to it by the AMD hardware) for all unassigned groups.
Thank you for clarifying. This seems specific to this software implementation,
and I missed that there was a shift from soft-RMIDs to soft-ABMC. If I remember
correctly this depends on undocumented hardware specific knowledge.
I consider software assignable a workaround to enable measuring
bandwidth reliably on a large number of groups on pre-ABMC AMD
hardware, or rather salvaging MBM on pre-ABMC hardware making use of
our users' effort to adapt to counter assignment in resctrl. We hope
no future implementations will choose to silently drop bandwidth
counts, so fingers crossed, the software implementation can be phased
out when these generations of AMD hardware are decommissioned.
That sounds ideal.
The MPAM specification natively supports (or requires) counter
assignment in hardware. From what I recall in the last of James'
prototypes I looked at, MBM was only supported if the implementation
provided as many bandwidth counters as there were possible monitoring
groups, so that it could assume a monitor IDs for every PARTID:PMG
combination.
Thank you for this insight.
What I would like to avoid is future complexity of needing a new mount/config
option that user space needs to use to select if a single "mbm_cntr_assignable"
is backed by hardware or software.
In my testing so far, automatically enabling counter assignment and
automatically allocating counters for all events in new groups works
well enough.
The only configuration I need is the ability to disable the automatic
counter allocation so that a userspace agent can have control of where
all the counters are assigned at all times. It's easy to implement
this as a simple flag if the user accepts that they need to manually
deallocate any automatically-allocated counters from groups created
before the flag was cleared.
The main semantic difference with SW assignments is that it is not
possible to assign counters to individual events. Because the
implementation is assigning RMIDs to groups, assignment results in all
events being counted.
I was considering introducing a boolean mbm_assign_events node to
indicate whether assigning individual events is supported. If true,
num_mbm_cntrs indicates the number of events which can be counted,
otherwise it indicates the number of groups to which counters can be
assigned and attempting to assign a single event is silently upgraded
to assigning counters to all events in the group.
How were you envisioning your users using the control file ("mbm_control")
in these scenarios? Does this file's interface even work for SW assignment
scenarios?
Users should expect consistent interface for "mbm_control" also.
It sounds to me that a potential "mbm_assign_events" will be false for SW
assignments. That would mean that "num_mbm_cntrs" will
contain the number of groups to which counters can be assigned?
Would user space be required to always enable all flags (enable all events) of
all domains to the same values ... or would enabling of one flag (one event)
in one domain automatically result in all flags (all events) enabled for all
domains ... or would enabling of one flag (one event) in one domain only appear
to user space to be enabled while in reality all flags/events are actually enabled?
I believe mbm_control should always accurately reflect which events
are being counted.
I agree.
The behavior as I've implemented today is:
# cat /sys/fs/resctrl/info/L3_MON/mbm_assign_events
0
# cat /sys/fs/resctrl/info/L3_MON/mbm_control
test//0=_;1=_;
//0=_;1=_;
# echo "test//1+l" > /sys/fs/resctrl/info/L3_MON/mbm_control
# cat /sys/fs/resctrl/info/L3_MON/mbm_control
test//0=_;1=tl;
//0=_;1=_;
# echo "test//1-t" > /sys/fs/resctrl/info/L3_MON/mbm_control
# cat /sys/fs/resctrl/info/L3_MON/mbm_control
test//0=_;1=_;
//0=_;1=_;
This highlights how there cannot be a generic/consistent interface between hardware
and software implementation. If resctrl implements something like above without any
other hints to user space then it will push complexity to user space since user space
would not know if setting one flag results in setting more than that flag, which may
force a user space implementation to always follow a write with a read that
needs to confirm what actually resulted from the write. Similarly, that removing a
flag impacts other flags needs to be clear without user space needing to "try and
see what happens".
It is not clear to me how to interpret the above example when it comes to the
RMID management though. If the RMID assignment is per group then I expected all
the domains of a group to have the same flag(s)?
However, If we don't expect to see these semantics in any other
implementation, these semantics could be implicit in the definition of
a SW assignable counter.
It is not clear to me how implementation differences between hardware
and software assignment can be hidden from user space. It is possible
to let user space enable individual events and then silently upgrade it
to all events. I see two options here, either "mbm_control" needs to
explicitly show this "silent upgrade" so that user space knows which
events are actually enabled, or "mbm_control" only shows flags/events enabled
from user space perspective. In the former scenario, this needs more
user space support since a generic user space cannot be confident which
flags are set after writing to "mbm_control". In the latter scenario,
meaning of "num_mbm_cntrs" becomes unclear since user space is expected
to rely on it to know which events can be enabled and if some are
actually "silently enabled" when user space still thinks it needs to be
enabled the number of available counters becomes vague.
It is not clear to me how to present hardware and software assignable
counters with a single consistent interface. Actually, what if the
"mbm_mode" is what distinguishes how counters are assigned instead of how
it is backed (hw vs sw)? What if, instead of "mbm_cntr_assignable" and
"mbm_cntr_sw_assignable" MBM modes the terms "mbm_cntr_event_assignable"
and "mbm_cntr_group_assignable" is used? Could that replace a
potential "mbm_assign_events" while also supporting user space in
interactions with "mbm_control"?
If I understand this correctly, is this a preference that the info
node be named differently if its value will have different units,
rather than a second node to indicate what the value of num_mbm_cntrs
actually means? This sounds reasonable to me.
Indeed. As you highlighted, user space may not need to know if
counters are backed by hardware or software, but user space needs to
know what to expect from (how to interact with) interface.
I think it's also important to note that in MPAM, the MBWU (memory
bandwidth usage) monitors don't have a concept of local versus total
bandwidth, so event assignment would likely not apply there either.
What the counted bandwidth actually represents is more implicit in the
monitor's position in the memory system in the particular
implementation. On a theoretical multi-socket system, resctrl would
require knowledge about the system's architecture to stitch together
the counts from different types of monitors to produce a local and
total value. I don't know if we'd program this SoC-specific knowledge
into the kernel to produce a unified MBM resource like we're
accustomed to now or if we'd present multiple MBM resources, each only
providing an mbm_total_bytes event. In this case, the counters would
have to be assigned separately in each MBM resource, especially if the
different MBM resources support a different number of counters.
"total" and "local" bandwidth is already in grey area after the
introduction of mbm_total_bytes_config/mbm_local_bytes_config where
user space could set values reported to not be constrained by the
"total" and "local" terms. We keep sticking with it though, even in
this implementation that uses the "t" and "l" flags, knowing that
what is actually monitored when "l" is set is just what the user
configured via mbm_local_bytes_config, which theoretically
can be "total" bandwidth.
Reinette
ps. I will be offline next week.
