Linaro-mm-sig

linaro-mm-sig@lists.linaro.org

16 participants
3178 discussions

Re: [RFC v2 02/13] dma-fence: Change signature of __dma_fence_is_later

by Christian König

On 5/9/25 17:33, Tvrtko Ursulin wrote: > With the goal of reducing the need for drivers to touch (and dereference) > fence->ops, we change the prototype of __dma_fence_is_later() to take > fence instead of fence->ops. > > Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin(a)igalia.com> Reviewed-by: Christian König <christian.koenig(a)amd.com> > --- > drivers/dma-buf/dma-fence-chain.c | 2 +- > drivers/dma-buf/sw_sync.c | 2 +- > drivers/gpu/drm/xe/xe_hw_fence.c | 2 +- > drivers/gpu/drm/xe/xe_sched_job.c | 14 ++++++++------ > include/linux/dma-fence.h | 9 ++++----- > 5 files changed, 15 insertions(+), 14 deletions(-) > > diff --git a/drivers/dma-buf/dma-fence-chain.c b/drivers/dma-buf/dma-fence-chain.c > index 9663ba1bb6ac..90424f23fd73 100644 > --- a/drivers/dma-buf/dma-fence-chain.c > +++ b/drivers/dma-buf/dma-fence-chain.c > @@ -252,7 +252,7 @@ void dma_fence_chain_init(struct dma_fence_chain *chain, > chain->prev_seqno = 0; > > /* Try to reuse the context of the previous chain node. */ > - if (prev_chain && __dma_fence_is_later(seqno, prev->seqno, prev->ops)) { > + if (prev_chain && __dma_fence_is_later(prev, seqno, prev->seqno)) { > context = prev->context; > chain->prev_seqno = prev->seqno; > } else { > diff --git a/drivers/dma-buf/sw_sync.c b/drivers/dma-buf/sw_sync.c > index 4f27ee93a00c..3c20f1d31cf5 100644 > --- a/drivers/dma-buf/sw_sync.c > +++ b/drivers/dma-buf/sw_sync.c > @@ -170,7 +170,7 @@ static bool timeline_fence_signaled(struct dma_fence *fence) > { > struct sync_timeline *parent = dma_fence_parent(fence); > > - return !__dma_fence_is_later(fence->seqno, parent->value, fence->ops); > + return !__dma_fence_is_later(fence, fence->seqno, parent->value); > } > > static void timeline_fence_set_deadline(struct dma_fence *fence, ktime_t deadline) > diff --git a/drivers/gpu/drm/xe/xe_hw_fence.c b/drivers/gpu/drm/xe/xe_hw_fence.c > index 0b4f12be3692..03eb8c6d1616 100644 > --- a/drivers/gpu/drm/xe/xe_hw_fence.c > +++ b/drivers/gpu/drm/xe/xe_hw_fence.c > @@ -165,7 +165,7 @@ static bool xe_hw_fence_signaled(struct dma_fence *dma_fence) > u32 seqno = xe_map_rd(xe, &fence->seqno_map, 0, u32); > > return dma_fence->error || > - !__dma_fence_is_later(dma_fence->seqno, seqno, dma_fence->ops); > + !__dma_fence_is_later(dma_fence, dma_fence->seqno, seqno); > } > > static bool xe_hw_fence_enable_signaling(struct dma_fence *dma_fence) > diff --git a/drivers/gpu/drm/xe/xe_sched_job.c b/drivers/gpu/drm/xe/xe_sched_job.c > index 1905ca590965..f0a6ce610948 100644 > --- a/drivers/gpu/drm/xe/xe_sched_job.c > +++ b/drivers/gpu/drm/xe/xe_sched_job.c > @@ -216,15 +216,17 @@ void xe_sched_job_set_error(struct xe_sched_job *job, int error) > > bool xe_sched_job_started(struct xe_sched_job *job) > { > + struct dma_fence *fence = dma_fence_chain_contained(job->fence); > struct xe_lrc *lrc = job->q->lrc[0]; > > - return !__dma_fence_is_later(xe_sched_job_lrc_seqno(job), > - xe_lrc_start_seqno(lrc), > - dma_fence_chain_contained(job->fence)->ops); > + return !__dma_fence_is_later(fence, > + xe_sched_job_lrc_seqno(job), > + xe_lrc_start_seqno(lrc)); > } > > bool xe_sched_job_completed(struct xe_sched_job *job) > { > + struct dma_fence *fence = dma_fence_chain_contained(job->fence); > struct xe_lrc *lrc = job->q->lrc[0]; > > /* > @@ -232,9 +234,9 @@ bool xe_sched_job_completed(struct xe_sched_job *job) > * parallel handshake is done. > */ > > - return !__dma_fence_is_later(xe_sched_job_lrc_seqno(job), > - xe_lrc_seqno(lrc), > - dma_fence_chain_contained(job->fence)->ops); > + return !__dma_fence_is_later(fence, > + xe_sched_job_lrc_seqno(job), > + xe_lrc_seqno(lrc)); > } > > void xe_sched_job_arm(struct xe_sched_job *job) > diff --git a/include/linux/dma-fence.h b/include/linux/dma-fence.h > index b12776883d14..48b5202c531d 100644 > --- a/include/linux/dma-fence.h > +++ b/include/linux/dma-fence.h > @@ -441,21 +441,20 @@ dma_fence_is_signaled(struct dma_fence *fence) > > /** > * __dma_fence_is_later - return if f1 is chronologically later than f2 > + * @fence: fence in whose context to do the comparison > * @f1: the first fence's seqno > * @f2: the second fence's seqno from the same context > - * @ops: dma_fence_ops associated with the seqno > * > * Returns true if f1 is chronologically later than f2. Both fences must be > * from the same context, since a seqno is not common across contexts. > */ > -static inline bool __dma_fence_is_later(u64 f1, u64 f2, > - const struct dma_fence_ops *ops) > +static inline bool __dma_fence_is_later(struct dma_fence *fence, u64 f1, u64 f2) > { > /* This is for backward compatibility with drivers which can only handle > * 32bit sequence numbers. Use a 64bit compare when the driver says to > * do so. > */ > - if (ops->use_64bit_seqno) > + if (fence->ops->use_64bit_seqno) > return f1 > f2; > > return (int)(lower_32_bits(f1) - lower_32_bits(f2)) > 0; > @@ -475,7 +474,7 @@ static inline bool dma_fence_is_later(struct dma_fence *f1, > if (WARN_ON(f1->context != f2->context)) > return false; > > - return __dma_fence_is_later(f1->seqno, f2->seqno, f1->ops); > + return __dma_fence_is_later(f1, f1->seqno, f2->seqno); > } > > /**

6 months, 2 weeks

Re: [RFC v2 01/13] drm/i915: Use provided dma_fence_is_chain

by Christian König

On 5/9/25 17:47, Matthew Brost wrote: > On Fri, May 09, 2025 at 04:33:40PM +0100, Tvrtko Ursulin wrote: >> Replace open-coded helper with the subsystem one. >> > > You probably can just send this one by itself as it good cleanup and > independent. > > Reviewed-by: Matthew Brost <matthew.brost(a)intel.com> Reviewed-by: Christian König <christian.koenig(a)amd.com> Any objections that I start to push those patches to drm-misc-next or do you want to take this one through the i915 branch? Regards, Christian. > >> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin(a)igalia.com> >> --- >> drivers/gpu/drm/i915/gem/i915_gem_wait.c | 7 +------ >> 1 file changed, 1 insertion(+), 6 deletions(-) >> >> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_wait.c b/drivers/gpu/drm/i915/gem/i915_gem_wait.c >> index 7127e90c1a8f..991666fd9f85 100644 >> --- a/drivers/gpu/drm/i915/gem/i915_gem_wait.c >> +++ b/drivers/gpu/drm/i915/gem/i915_gem_wait.c >> @@ -106,11 +106,6 @@ static void fence_set_priority(struct dma_fence *fence, >> rcu_read_unlock(); >> } >> >> -static inline bool __dma_fence_is_chain(const struct dma_fence *fence) >> -{ >> - return fence->ops == &dma_fence_chain_ops; >> -} >> - >> void i915_gem_fence_wait_priority(struct dma_fence *fence, >> const struct i915_sched_attr *attr) >> { >> @@ -126,7 +121,7 @@ void i915_gem_fence_wait_priority(struct dma_fence *fence, >> >> for (i = 0; i < array->num_fences; i++) >> fence_set_priority(array->fences[i], attr); >> - } else if (__dma_fence_is_chain(fence)) { >> + } else if (dma_fence_is_chain(fence)) { >> struct dma_fence *iter; >> >> /* The chain is ordered; if we boost the last, we boost all */ >> -- >> 2.48.0 >>

6 months, 2 weeks

Re: [PATCH bpf-next v4 5/5] selftests/bpf: Add test for open coded dmabuf_iter

by T.J. Mercier

On Fri, May 9, 2025 at 2:58 PM Song Liu <song(a)kernel.org> wrote: > > On Fri, May 9, 2025 at 2:43 PM T.J. Mercier <tjmercier(a)google.com> wrote: > > > [...] > > > > > > Personally, I would prefer we just merge all the logic of > > > create_udmabuf() and create_sys_heap_dmabuf() > > > into create_test_buffers(). > > > > That's a lot of different stuff to put in one place. How about > > returning file descriptors from the buffer create functions while > > having them clean up after themselves: > > I do like this version better. Some nitpicks though. > > > > > -static int memfd, udmabuf; > > +static int udmabuf; > > About this, and ... > > > static const char udmabuf_test_buffer_name[DMA_BUF_NAME_LEN] = > > "udmabuf_test_buffer_for_iter"; > > static size_t udmabuf_test_buffer_size; > > static int sysheap_dmabuf; > > static const char sysheap_test_buffer_name[DMA_BUF_NAME_LEN] = > > "sysheap_test_buffer_for_iter"; > > static size_t sysheap_test_buffer_size; > > > > -static int create_udmabuf(int map_fd) > > +static int create_udmabuf(void) > > { > > struct udmabuf_create create; > > - int dev_udmabuf; > > - bool f = false; > > + int dev_udmabuf, memfd, udmabuf; > .. here. > > It is not ideal to have a global udmabuf and a local udmabuf. > If we want the global version, let's rename the local one. Ok let me change up the name of the aliasing variable to local_udmabuf. > [...] > > > > > static int create_test_buffers(int map_fd) > > { > > - int ret; > > + bool f = false; > > + > > + udmabuf = create_udmabuf(); > > + sysheap_dmabuf = create_sys_heap_dmabuf(); > > > > - ret = create_udmabuf(map_fd); > > - if (ret) > > - return ret; > > + if (udmabuf < 0 || sysheap_dmabuf < 0) > > + return -1; > > We also need destroy_test_buffers() on the error path here, > or at the caller. The caller does currently check to decide if it should bother running the tests or not, and calls destroy_test_buffers() if not. > > - return create_sys_heap_dmabuf(map_fd); > > + return bpf_map_update_elem(map_fd, udmabuf_test_buffer_name, > > &f, BPF_ANY) || > > + bpf_map_update_elem(map_fd, sysheap_test_buffer_name, > > &f, BPF_ANY); > > } > > > > static void destroy_test_buffers(void) > > { > > close(udmabuf); > > - close(memfd); > > close(sysheap_dmabuf); > > For the two global fds, let's reset them to -1 right after close(). > > Thanks, > Song Will do, thanks.

6 months, 2 weeks

Re: [PATCH bpf-next v4 5/5] selftests/bpf: Add test for open coded dmabuf_iter

by T.J. Mercier

On Fri, May 9, 2025 at 11:46 AM Song Liu <song(a)kernel.org> wrote: > > On Thu, May 8, 2025 at 11:21 AM T.J. Mercier <tjmercier(a)google.com> wrote: > > > > Use the same test buffers as the traditional iterator and a new BPF map > > to verify the test buffers can be found with the open coded dmabuf > > iterator. > > The way we split 4/5 and 5/5 makes the code tricker to follow. I guess > the motivation is to back port default iter along to older kernels. But I > think we can still make the code cleaner. > > > > > Signed-off-by: T.J. Mercier <tjmercier(a)google.com> > > --- > [...] > > > > > -static int create_udmabuf(void) > > +static int create_udmabuf(int map_fd) > > { > > struct udmabuf_create create; > > int dev_udmabuf; > > + bool f = false; > > > > udmabuf_test_buffer_size = 10 * getpagesize(); > > > > @@ -63,10 +64,10 @@ static int create_udmabuf(void) > > if (!ASSERT_OK(ioctl(udmabuf, DMA_BUF_SET_NAME_B, udmabuf_test_buffer_name), "name")) > > return 1; > > > > - return 0; > > + return bpf_map_update_elem(map_fd, udmabuf_test_buffer_name, &f, BPF_ANY); > > We don't really need this bpf_map_update_elem() inside > create_udmabuf(), right? > > > } > > > > -static int create_sys_heap_dmabuf(void) > > +static int create_sys_heap_dmabuf(int map_fd) > > { > > sysheap_test_buffer_size = 20 * getpagesize(); > > > > @@ -77,6 +78,7 @@ static int create_sys_heap_dmabuf(void) > > .heap_flags = 0, > > }; > > int heap_fd, ret; > > + bool f = false; > > > > if (!ASSERT_LE(sizeof(sysheap_test_buffer_name), DMA_BUF_NAME_LEN, "NAMETOOLONG")) > > return 1; > > @@ -95,18 +97,18 @@ static int create_sys_heap_dmabuf(void) > > if (!ASSERT_OK(ioctl(sysheap_dmabuf, DMA_BUF_SET_NAME_B, sysheap_test_buffer_name), "name")) > > return 1; > > > > - return 0; > > + return bpf_map_update_elem(map_fd, sysheap_test_buffer_name, &f, BPF_ANY); > > Same for this bpf_map_update_elem(), we can call this directly from > create_test_buffers(). > > > } > > > > -static int create_test_buffers(void) > > +static int create_test_buffers(int map_fd) > > { > > int ret; > > > > - ret = create_udmabuf(); > > + ret = create_udmabuf(map_fd); > > if (ret) > > return ret; > > > > - return create_sys_heap_dmabuf(); > > + return create_sys_heap_dmabuf(map_fd); > > Personally, I would prefer we just merge all the logic of > create_udmabuf() and create_sys_heap_dmabuf() > into create_test_buffers(). That's a lot of different stuff to put in one place. How about returning file descriptors from the buffer create functions while having them clean up after themselves: -static int memfd, udmabuf; +static int udmabuf; static const char udmabuf_test_buffer_name[DMA_BUF_NAME_LEN] = "udmabuf_test_buffer_for_iter"; static size_t udmabuf_test_buffer_size; static int sysheap_dmabuf; static const char sysheap_test_buffer_name[DMA_BUF_NAME_LEN] = "sysheap_test_buffer_for_iter"; static size_t sysheap_test_buffer_size; -static int create_udmabuf(int map_fd) +static int create_udmabuf(void) { struct udmabuf_create create; - int dev_udmabuf; - bool f = false; + int dev_udmabuf, memfd, udmabuf; udmabuf_test_buffer_size = 10 * getpagesize(); if (!ASSERT_LE(sizeof(udmabuf_test_buffer_name), DMA_BUF_NAME_LEN, "NAMETOOLONG")) - return 1; + return -1; memfd = memfd_create("memfd_test", MFD_ALLOW_SEALING); if (!ASSERT_OK_FD(memfd, "memfd_create")) - return 1; + return -1; if (!ASSERT_OK(ftruncate(memfd, udmabuf_test_buffer_size), "ftruncate")) - return 1; + goto close_memfd; if (!ASSERT_OK(fcntl(memfd, F_ADD_SEALS, F_SEAL_SHRINK), "seal")) - return 1; + goto close_memfd; dev_udmabuf = open("/dev/udmabuf", O_RDONLY); if (!ASSERT_OK_FD(dev_udmabuf, "open udmabuf")) - return 1; + goto close_memfd; create.memfd = memfd; create.flags = UDMABUF_FLAGS_CLOEXEC; @@ -59,15 +58,21 @@ static int create_udmabuf(int map_fd) udmabuf = ioctl(dev_udmabuf, UDMABUF_CREATE, &create); close(dev_udmabuf); if (!ASSERT_OK_FD(udmabuf, "udmabuf_create")) - return 1; + goto close_memfd; if (!ASSERT_OK(ioctl(udmabuf, DMA_BUF_SET_NAME_B, udmabuf_test_buffer_name), "name")) - return 1; + goto close_udmabuf; + + return udmabuf; - return bpf_map_update_elem(map_fd, udmabuf_test_buffer_name, &f, BPF_ANY); +close_udmabuf: + close(udmabuf); +close_memfd: + close(memfd); + return -1; } -static int create_sys_heap_dmabuf(int map_fd) +static int create_sys_heap_dmabuf(void) { sysheap_test_buffer_size = 20 * getpagesize(); @@ -78,43 +83,46 @@ static int create_sys_heap_dmabuf(int map_fd) .heap_flags = 0, }; int heap_fd, ret; - bool f = false; if (!ASSERT_LE(sizeof(sysheap_test_buffer_name), DMA_BUF_NAME_LEN, "NAMETOOLONG")) - return 1; + return -1; heap_fd = open("/dev/dma_heap/system", O_RDONLY); if (!ASSERT_OK_FD(heap_fd, "open dma heap")) - return 1; + return -1; ret = ioctl(heap_fd, DMA_HEAP_IOCTL_ALLOC, &data); close(heap_fd); if (!ASSERT_OK(ret, "syheap alloc")) - return 1; + return -1; - sysheap_dmabuf = data.fd; + if (!ASSERT_OK(ioctl(data.fd, DMA_BUF_SET_NAME_B, sysheap_test_buffer_name), "name")) + goto close_sysheap_dmabuf; - if (!ASSERT_OK(ioctl(sysheap_dmabuf, DMA_BUF_SET_NAME_B, sysheap_test_buffer_name), "name")) - return 1; + return data.fd; - return bpf_map_update_elem(map_fd, sysheap_test_buffer_name, &f, BPF_ANY); +close_sysheap_dmabuf: + close(data.fd); + return -1; } static int create_test_buffers(int map_fd) { - int ret; + bool f = false; + + udmabuf = create_udmabuf(); + sysheap_dmabuf = create_sys_heap_dmabuf(); - ret = create_udmabuf(map_fd); - if (ret) - return ret; + if (udmabuf < 0 || sysheap_dmabuf < 0) + return -1; - return create_sys_heap_dmabuf(map_fd); + return bpf_map_update_elem(map_fd, udmabuf_test_buffer_name, &f, BPF_ANY) || + bpf_map_update_elem(map_fd, sysheap_test_buffer_name, &f, BPF_ANY); } static void destroy_test_buffers(void) { close(udmabuf); - close(memfd); close(sysheap_dmabuf); }

6 months, 2 weeks

Re: [RFC PATCH 00/12] Private MMIO support for private assigned dev

by Jason Gunthorpe

On Sat, May 10, 2025 at 12:28:48AM +0800, Xu Yilun wrote: > On Fri, May 09, 2025 at 07:12:46PM +0800, Xu Yilun wrote: > > On Fri, May 09, 2025 at 01:04:58PM +1000, Alexey Kardashevskiy wrote: > > > Ping? > > > > Sorry for late reply from vacation. > > > > > Also, since there is pushback on 01/12 "dma-buf: Introduce dma_buf_get_pfn_unlocked() kAPI", what is the plan now? Thanks, > > > > As disscussed in the thread, this kAPI is not well considered but IIUC > > the concept of "importer mapping" is still valid. We need more > > investigation about all the needs - P2P, CC memory, private bus > > channel, and work out a formal API. > > > > However in last few months I'm focusing on high level TIO flow - TSM > > framework, IOMMUFD based bind/unbind, so no much progress here and is > > still using this temporary kAPI. But as long as "importer mapping" is > > alive, the dmabuf fd for KVM is still valid and we could enable TIO > > based on that. > > Oh I forgot to mention I moved the dmabuf creation from VFIO to IOMMUFD > recently, the IOCTL is against iommufd_device. I'm surprised by this.. iommufd shouldn't be doing PCI stuff, it is just about managing the translation control of the device. > According to Jason's > opinion [1], TSM bind/unbind should be called against iommufd_device, > then I need to do the same for dmabuf. This is because Intel TDX > Connect enforces a specific operation sequence between TSM unbind & MMIO > unmap: > > 1. STOP TDI via TDISP message STOP_INTERFACE > 2. Private MMIO unmap from Secure EPT > 3. Trusted Device Context Table cleanup for the TDI > 4. TDI ownership reclaim and metadata free So your issue is you need to shoot down the dmabuf during vPCI device destruction? VFIO also needs to shoot down the MMIO during things like FLR I don't think moving to iommufd really fixes it, it sounds like you need more coordination between the two parts?? Jason

6 months, 2 weeks

Re: [PATCH bpf-next v4 4/5] selftests/bpf: Add test for dmabuf_iter

by T.J. Mercier

On Thu, May 8, 2025 at 5:36 PM Song Liu <song(a)kernel.org> wrote: > > On Thu, May 8, 2025 at 11:20 AM T.J. Mercier <tjmercier(a)google.com> wrote: > [...] > > diff --git a/tools/testing/selftests/bpf/prog_tests/dmabuf_iter.c b/tools/testing/selftests/bpf/prog_tests/dmabuf_iter.c > > new file mode 100644 > > index 000000000000..35745f4ce0f8 > > --- /dev/null > > +++ b/tools/testing/selftests/bpf/prog_tests/dmabuf_iter.c > > @@ -0,0 +1,224 @@ > > +// SPDX-License-Identifier: GPL-2.0 > > +/* Copyright (c) 2025 Google */ > > + > > +#include <test_progs.h> > > +#include <bpf/libbpf.h> > > +#include <bpf/btf.h> > > +#include "dmabuf_iter.skel.h" > > + > > +#include <fcntl.h> > > +#include <stdbool.h> > > +#include <stdio.h> > > +#include <stdlib.h> > > +#include <string.h> > > +#include <sys/ioctl.h> > > +#include <sys/mman.h> > > +#include <unistd.h> > > + > > +#include <linux/dma-buf.h> > > +#include <linux/dma-heap.h> > > +#include <linux/udmabuf.h> > > + > > +static int memfd, udmabuf; > > Global fds are weird. AFAICT, we don't really need them > to be global? If we really need them to be global, please > initialize them to -1, just in case we close(0) by accident. Hmm, no we don't really need them to be global but I didn't really want to pass all these variables around to all the setup and test functions. The fd lifetimes are nearly the whole program lifetime anyways, and just need to exist without actually being used for anything. I'll add the -1 initialization as you suggest. If udmabuf creation failed, we would have done a close(0) in destroy_test_buffers() on the sysheap_dmabuf fd. > > +static const char udmabuf_test_buffer_name[DMA_BUF_NAME_LEN] = "udmabuf_test_buffer_for_iter"; > > +static size_t udmabuf_test_buffer_size; > > +static int sysheap_dmabuf; > > +static const char sysheap_test_buffer_name[DMA_BUF_NAME_LEN] = "sysheap_test_buffer_for_iter"; > > +static size_t sysheap_test_buffer_size;

6 months, 2 weeks

Re: [PATCH bpf-next v4 3/5] bpf: Add open coded dmabuf iterator

by T.J. Mercier

On Thu, May 8, 2025 at 5:28 PM Song Liu <song(a)kernel.org> wrote: > > On Thu, May 8, 2025 at 11:20 AM T.J. Mercier <tjmercier(a)google.com> wrote: > > > > This open coded iterator allows for more flexibility when creating BPF > > programs. It can support output in formats other than text. With an open > > coded iterator, a single BPF program can traverse multiple kernel data > > structures (now including dmabufs), allowing for more efficient analysis > > of kernel data compared to multiple reads from procfs, sysfs, or > > multiple traditional BPF iterator invocations. > > > > Signed-off-by: T.J. Mercier <tjmercier(a)google.com> > > Acked-by: Song Liu <song(a)kernel.org> > > With one nitpick below: > > > --- > > kernel/bpf/dmabuf_iter.c | 47 ++++++++++++++++++++++++++++++++++++++++ > > kernel/bpf/helpers.c | 5 +++++ > > 2 files changed, 52 insertions(+) > > > > diff --git a/kernel/bpf/dmabuf_iter.c b/kernel/bpf/dmabuf_iter.c > > index 96b4ba7f0b2c..8049bdbc9efc 100644 > > --- a/kernel/bpf/dmabuf_iter.c > > +++ b/kernel/bpf/dmabuf_iter.c > > @@ -100,3 +100,50 @@ static int __init dmabuf_iter_init(void) > > } > > > > late_initcall(dmabuf_iter_init); > > + > > +struct bpf_iter_dmabuf { > > + /* opaque iterator state; having __u64 here allows to preserve correct > > + * alignment requirements in vmlinux.h, generated from BTF > > + */ > > nit: comment style. Added a leading /* (This is copied from task_iter.c, which currently has the same style.) > > + __u64 __opaque[1]; > > +} __aligned(8); > > + > > +/* Non-opaque version of bpf_iter_dmabuf */ > > +struct bpf_iter_dmabuf_kern { > > + struct dma_buf *dmabuf; > > +} __aligned(8); > > + > [...]

6 months, 2 weeks

Re: [PATCH bpf-next v4 2/5] bpf: Add dmabuf iterator

by T.J. Mercier

On Thu, May 8, 2025 at 5:27 PM Song Liu <song(a)kernel.org> wrote: > > On Thu, May 8, 2025 at 11:20 AM T.J. Mercier <tjmercier(a)google.com> wrote: > > > > The dmabuf iterator traverses the list of all DMA buffers. > > > > DMA buffers are refcounted through their associated struct file. A > > reference is taken on each buffer as the list is iterated to ensure each > > buffer persists for the duration of the bpf program execution without > > holding the list mutex. > > > > Signed-off-by: T.J. Mercier <tjmercier(a)google.com> > > Reviewed-by: Christian König <christian.koenig(a)amd.com> > > Acked-by: Song Liu <song(a)kernel.org> > > With one nitpick below. Thanks! > > --- > [...] > > diff --git a/include/linux/dma-buf.h b/include/linux/dma-buf.h > > index 8ff4add71f88..7af2ea839f58 100644 > > --- a/include/linux/dma-buf.h > > +++ b/include/linux/dma-buf.h > > @@ -634,4 +634,6 @@ int dma_buf_vmap(struct dma_buf *dmabuf, struct iosys_map *map); > > void dma_buf_vunmap(struct dma_buf *dmabuf, struct iosys_map *map); > > int dma_buf_vmap_unlocked(struct dma_buf *dmabuf, struct iosys_map *map); > > void dma_buf_vunmap_unlocked(struct dma_buf *dmabuf, struct iosys_map *map); > > +struct dma_buf *dma_buf_iter_begin(void); > > +struct dma_buf *dma_buf_iter_next(struct dma_buf *dmbuf); > > #endif /* __DMA_BUF_H__ */ > > diff --git a/kernel/bpf/Makefile b/kernel/bpf/Makefile > > index 70502f038b92..3a335c50e6e3 100644 > > --- a/kernel/bpf/Makefile > > +++ b/kernel/bpf/Makefile > > @@ -53,6 +53,9 @@ obj-$(CONFIG_BPF_SYSCALL) += relo_core.o > > obj-$(CONFIG_BPF_SYSCALL) += btf_iter.o > > obj-$(CONFIG_BPF_SYSCALL) += btf_relocate.o > > obj-$(CONFIG_BPF_SYSCALL) += kmem_cache_iter.o > > +ifeq ($(CONFIG_DMA_SHARED_BUFFER),y) > > +obj-$(CONFIG_BPF_SYSCALL) += dmabuf_iter.o > > +endif > > > > CFLAGS_REMOVE_percpu_freelist.o = $(CC_FLAGS_FTRACE) > > CFLAGS_REMOVE_bpf_lru_list.o = $(CC_FLAGS_FTRACE) > > diff --git a/kernel/bpf/dmabuf_iter.c b/kernel/bpf/dmabuf_iter.c > > new file mode 100644 > > index 000000000000..96b4ba7f0b2c > > --- /dev/null > > +++ b/kernel/bpf/dmabuf_iter.c > > @@ -0,0 +1,102 @@ > > +// SPDX-License-Identifier: GPL-2.0-only > > +/* Copyright (c) 2025 Google LLC */ > > +#include <linux/bpf.h> > > +#include <linux/btf_ids.h> > > +#include <linux/dma-buf.h> > > +#include <linux/kernel.h> > > +#include <linux/seq_file.h> > > + > > +BTF_ID_LIST_SINGLE(bpf_dmabuf_btf_id, struct, dma_buf) > > +DEFINE_BPF_ITER_FUNC(dmabuf, struct bpf_iter_meta *meta, struct dma_buf *dmabuf) > > nit: It is better to move these two lines later, to where they > are about to be used. I've moved them both to just before dmabuf_iter_init() farther down.

6 months, 2 weeks

[PATCH bpf-next v4 0/5] Replace CONFIG_DMABUF_SYSFS_STATS with BPF

by T.J. Mercier

Until CONFIG_DMABUF_SYSFS_STATS was added [1] it was only possible to perform per-buffer accounting with debugfs which is not suitable for production environments. Eventually we discovered the overhead with per-buffer sysfs file creation/removal was significantly impacting allocation and free times, and exacerbated kernfs lock contention. [2] dma_buf_stats_setup() is responsible for 39% of single-page buffer creation duration, or 74% of single-page dma_buf_export() duration when stressing dmabuf allocations and frees. I prototyped a change from per-buffer to per-exporter statistics with a RCU protected list of exporter allocations that accommodates most (but not all) of our use-cases and avoids almost all of the sysfs overhead. While that adds less overhead than per-buffer sysfs, and less even than the maintenance of the dmabuf debugfs_list, it's still *additional* overhead on top of the debugfs_list and doesn't give us per-buffer info. This series uses the existing dmabuf debugfs_list to implement a BPF dmabuf iterator, which adds no overhead to buffer allocation/free and provides per-buffer info. The list has been moved outside of CONFIG_DEBUG_FS scope so that it is always populated. The BPF program loaded by userspace that extracts per-buffer information gets to define its own interface which avoids the lack of ABI stability with debugfs. This will allow us to replace our use of CONFIG_DMABUF_SYSFS_STATS, and the plan is to remove it from the kernel after the next longterm stable release. [1] https://lore.kernel.org/linux-media/20201210044400.1080308-1-hridya@google.… [2] https://lore.kernel.org/all/20220516171315.2400578-1-tjmercier@google.com v1: https://lore.kernel.org/all/20250414225227.3642618-1-tjmercier@google.com v1 -> v2: Make the DMA buffer list independent of CONFIG_DEBUG_FS per Christian König Add CONFIG_DMA_SHARED_BUFFER check to kernel/bpf/Makefile per kernel test robot Use BTF_ID_LIST_SINGLE instead of BTF_ID_LIST_GLOBAL_SINGLE per Song Liu Fixup comment style, mixing code/declarations, and use ASSERT_OK_FD in selftest per Song Liu Add BPF_ITER_RESCHED feature to bpf_dmabuf_reg_info per Alexei Starovoitov Add open-coded iterator and selftest per Alexei Starovoitov Add a second test buffer from the system dmabuf heap to selftests Use the BPF program we'll use in production for selftest per Alexei Starovoitov https://r.android.com/c/platform/system/bpfprogs/+/3616123/2/dmabufIter.c https://r.android.com/c/platform/system/memory/libmeminfo/+/3614259/1/libdm… v2: https://lore.kernel.org/all/20250504224149.1033867-1-tjmercier@google.com v2 -> v3: Rebase onto bpf-next/master Move get_next_dmabuf() into drivers/dma-buf/dma-buf.c, along with the new get_first_dmabuf(). This avoids having to expose the dmabuf list and mutex to the rest of the kernel, and keeps the dmabuf mutex operations near each other in the same file. (Christian König) Add Christian's RB to dma-buf: Rename debugfs symbols Drop RFC: dma-buf: Remove DMA-BUF statistics v3: https://lore.kernel.org/all/20250507001036.2278781-1-tjmercier@google.com v3 -> v4: Fix selftest BPF program comment style (not kdoc) per Alexei Starovoitov Fix dma-buf.c kdoc comment style per Alexei Starovoitov Rename get_first_dmabuf / get_next_dmabuf to dma_buf_iter_begin / dma_buf_iter_next per Christian König Add Christian's RB to bpf: Add dmabuf iterator T.J. Mercier (5): dma-buf: Rename debugfs symbols bpf: Add dmabuf iterator bpf: Add open coded dmabuf iterator selftests/bpf: Add test for dmabuf_iter selftests/bpf: Add test for open coded dmabuf_iter drivers/dma-buf/dma-buf.c | 98 +++++-- include/linux/dma-buf.h | 4 +- kernel/bpf/Makefile | 3 + kernel/bpf/dmabuf_iter.c | 149 ++++++++++ kernel/bpf/helpers.c | 5 + .../testing/selftests/bpf/bpf_experimental.h | 5 + tools/testing/selftests/bpf/config | 3 + .../selftests/bpf/prog_tests/dmabuf_iter.c | 258 ++++++++++++++++++ .../testing/selftests/bpf/progs/dmabuf_iter.c | 91 ++++++ 9 files changed, 594 insertions(+), 22 deletions(-) create mode 100644 kernel/bpf/dmabuf_iter.c create mode 100644 tools/testing/selftests/bpf/prog_tests/dmabuf_iter.c create mode 100644 tools/testing/selftests/bpf/progs/dmabuf_iter.c base-commit: 43745d11bfd9683abdf08ad7a5cc403d6a9ffd15 -- 2.49.0.1015.ga840276032-goog

6 months, 2 weeks

Re: [PATCH] dma-buf/sw-sync: Remove unused debug code

by Christian König

On 5/6/25 01:38, linux(a)treblig.org wrote: > From: "Dr. David Alan Gilbert" <linux(a)treblig.org> > > sync_file_debug_add() and sync_file_debug_remove() have been unused > since 2016's > commit d4cab38e153d ("staging/android: prepare sync_file for de-staging") > > Remove them. > > Since sync_file_debug_add was the only thing to add to > sync_file_list_head, the code that dumps it in part of > sync_info_debugfs_show can be removed, and the declaration of > the list and it's associated lock can be removed. > (The 'fences:\n...' marker in that debugfs file is left in > so as not to change the output) > > That leaves the sync_print_sync_file() helper unused, and > is thus removed. > > Signed-off-by: Dr. David Alan Gilbert <linux(a)treblig.org> I've added my reviewed-by and pushed it into drm-misc-next for upstreaming. Thanks, Christian. > --- > drivers/dma-buf/sync_debug.c | 49 ------------------------------------ > drivers/dma-buf/sync_debug.h | 2 -- > 2 files changed, 51 deletions(-) > > diff --git a/drivers/dma-buf/sync_debug.c b/drivers/dma-buf/sync_debug.c > index 237bce21d1e7..a9c3312dc85d 100644 > --- a/drivers/dma-buf/sync_debug.c > +++ b/drivers/dma-buf/sync_debug.c > @@ -12,8 +12,6 @@ static struct dentry *dbgfs; > > static LIST_HEAD(sync_timeline_list_head); > static DEFINE_SPINLOCK(sync_timeline_list_lock); > -static LIST_HEAD(sync_file_list_head); > -static DEFINE_SPINLOCK(sync_file_list_lock); > > void sync_timeline_debug_add(struct sync_timeline *obj) > { > @@ -33,24 +31,6 @@ void sync_timeline_debug_remove(struct sync_timeline *obj) > spin_unlock_irqrestore(&sync_timeline_list_lock, flags); > } > > -void sync_file_debug_add(struct sync_file *sync_file) > -{ > - unsigned long flags; > - > - spin_lock_irqsave(&sync_file_list_lock, flags); > - list_add_tail(&sync_file->sync_file_list, &sync_file_list_head); > - spin_unlock_irqrestore(&sync_file_list_lock, flags); > -} > - > -void sync_file_debug_remove(struct sync_file *sync_file) > -{ > - unsigned long flags; > - > - spin_lock_irqsave(&sync_file_list_lock, flags); > - list_del(&sync_file->sync_file_list); > - spin_unlock_irqrestore(&sync_file_list_lock, flags); > -} > - > static const char *sync_status_str(int status) > { > if (status < 0) > @@ -118,26 +98,6 @@ static void sync_print_obj(struct seq_file *s, struct sync_timeline *obj) > spin_unlock(&obj->lock); > } > > -static void sync_print_sync_file(struct seq_file *s, > - struct sync_file *sync_file) > -{ > - char buf[128]; > - int i; > - > - seq_printf(s, "[%p] %s: %s\n", sync_file, > - sync_file_get_name(sync_file, buf, sizeof(buf)), > - sync_status_str(dma_fence_get_status(sync_file->fence))); > - > - if (dma_fence_is_array(sync_file->fence)) { > - struct dma_fence_array *array = to_dma_fence_array(sync_file->fence); > - > - for (i = 0; i < array->num_fences; ++i) > - sync_print_fence(s, array->fences[i], true); > - } else { > - sync_print_fence(s, sync_file->fence, true); > - } > -} > - > static int sync_info_debugfs_show(struct seq_file *s, void *unused) > { > struct list_head *pos; > @@ -157,15 +117,6 @@ static int sync_info_debugfs_show(struct seq_file *s, void *unused) > > seq_puts(s, "fences:\n--------------\n"); > > - spin_lock_irq(&sync_file_list_lock); > - list_for_each(pos, &sync_file_list_head) { > - struct sync_file *sync_file = > - container_of(pos, struct sync_file, sync_file_list); > - > - sync_print_sync_file(s, sync_file); > - seq_putc(s, '\n'); > - } > - spin_unlock_irq(&sync_file_list_lock); > return 0; > } > > diff --git a/drivers/dma-buf/sync_debug.h b/drivers/dma-buf/sync_debug.h > index a1bdd62efccd..02af347293d0 100644 > --- a/drivers/dma-buf/sync_debug.h > +++ b/drivers/dma-buf/sync_debug.h > @@ -68,7 +68,5 @@ extern const struct file_operations sw_sync_debugfs_fops; > > void sync_timeline_debug_add(struct sync_timeline *obj); > void sync_timeline_debug_remove(struct sync_timeline *obj); > -void sync_file_debug_add(struct sync_file *fence); > -void sync_file_debug_remove(struct sync_file *fence); > > #endif /* _LINUX_SYNC_H */

6 months, 2 weeks

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

Linaro-mm-sig