From dd05add7b2b6004c5fa3a1f276f56d8668f86b65 Mon Sep 17 00:00:00 2001 From: Paulo Alcantara Date: Sat, 26 Sep 2026 19:20:30 -0300 Subject: netfs: zero the tail of a short DIO/unbuffered read The buffered read collector zero-fills the tail of a short read that stops below the inode's i_size (netfs_clear_unread()), so a read that races an extending write still returns the expected number of bytes. The non-buffered collector path does no such thing: it just records how much was transferred. Add the same zero-fill for the non-buffered case, gated on NETFS_SREQ_CLEAR_TAIL: a subreq's source sets that flag when a short result from it is known to be safe to treat as a hole, as opposed to NETFS_SREQ_HIT_EOF, which means the read genuinely ran off the end of the file and should be reported short as-is. Only CLEAR_TAIL should zero-fill here; a real EOF must stay a real short read. No source currently sets CLEAR_TAIL on an unbuffered/DIO subrequest, so this is inert on its own -- a following change teaches cifs to set it in the one case that needs it. Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item") Reviewed-by: David Howells Reviewed-by: Namjae Jeon Signed-off-by: Paulo Alcantara Cc: Christian Brauner Cc: Matthew Wilcox Cc: Ronnie Sahlberg Cc: Shyam Prasad N Cc: Tom Talpey Cc: Bharath SM Cc: stable@vger.kernel.org --- Documentation/gpu/drm-compute.rst | 54 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 54 insertions(+) create mode 100644 Documentation/gpu/drm-compute.rst (limited to 'Documentation/gpu/drm-compute.rst') diff --git a/Documentation/gpu/drm-compute.rst b/Documentation/gpu/drm-compute.rst new file mode 100644 index 000000000..f90c3e63a --- /dev/null +++ b/Documentation/gpu/drm-compute.rst @@ -0,0 +1,54 @@ +================================== +Long running workloads and compute +================================== + +Long running workloads (compute) are workloads that will not complete in 10 +seconds. (The time let the user wait before he reaches for the power button). +This means that other techniques need to be used to manage those workloads, +that cannot use fences. + +Some hardware may schedule compute jobs, and have no way to pre-empt them, or +have their memory swapped out from them. Or they simply want their workload +not to be preempted or swapped out at all. + +This means that it differs from what is described in driver-api/dma-buf.rst. + +As with normal compute jobs, dma-fence may not be used at all. In this case, +not even to force preemption. The driver with is simply forced to unmap a BO +from the long compute job's address space on unbind immediately, not even +waiting for the workload to complete. Effectively this terminates the workload +when there is no hardware support to recover. + +Since this is undesirable, there need to be mitigations to prevent a workload +from being terminated. There are several possible approach, all with their +advantages and drawbacks. + +The first approach you will likely try is to pin all buffers used by compute. +This guarantees that the job will run uninterrupted, but also allows a very +denial of service attack by pinning as much memory as possible, hogging the +all GPU memory, and possibly a huge chunk of CPU memory. + +A second approach that will work slightly better on its own is adding an option +not to evict when creating a new job (any kind). If all of userspace opts in +to this flag, it would prevent cooperating userspace from forced terminating +older compute jobs to start a new one. + +If job preemption and recoverable pagefaults are not available, those are the +only approaches possible. So even with those, you want a separate way of +controlling resources. The standard kernel way of doing so is cgroups. + +This creates a third option, using cgroups to prevent eviction. Both GPU and +driver-allocated CPU memory would be accounted to the correct cgroup, and +eviction would be made cgroup aware. This allows the GPU to be partitioned +into cgroups, that will allow jobs to run next to each other without +interference. + +The interface to the cgroup would be similar to the current CPU memory +interface, with similar semantics for min/low/high/max, if eviction can +be made cgroup aware. + +What should be noted is that each memory region (tiled memory for example) +should have its own accounting. + +The key is set to the regionid set by the driver, for example "tile0". +For the value of $card, we use drmGetUnique(). -- cgit v1.3.1