Wb on itr experiments #23

michalQb · 2024-02-20T16:11:27Z

No description provided.

After commit 5027ec1 ("net: page_pool: split the page_pool_params into fast and slow") that made &page_pool contain only "hot" params at the start, cacheline boundary chops frag API fields group in the middle again. To not bother with this each time fast params get expanded or shrunk, let's just align them to `4 * sizeof(long)`, the closest upper pow-2 to their actual size (2 longs + 2 ints). This ensures 16-byte alignment for the 32-bit architectures and 32-byte alignment for the 64-bit ones, excluding unnecessary false-sharing. Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Quite often, NIC devices do not need dma_sync operations on x86_64 at least. Indeed, when dev_is_dma_coherent(dev) is true and dev_use_swiotlb(dev) is false, iommu_dma_sync_single_for_cpu() and friends do nothing. However, indirectly calling them when CONFIG_RETPOLINE=y consumes about 10% of cycles on a cpu receiving packets from softirq at ~100Gbit rate, as shown in [1] Even if/when CONFIG_RETPOLINE is not set, there is a cost of about 3%. Add dev->skip_dma_sync boolean which is set during the device initialization depending on the setup: dev_is_dma_coherent() for direct DMA, !(sync_single_for_device || sync_single_for_cpu) or positive result from the new callback, dma_map_ops::can_skip_sync for non-NULL DMA ops. Then later, if/when swiotlb is used for the first time, the flag is turned off, from swiotlb_tbl_map_single(). perf profile before the patch: 18.53% [kernel] [k] gq_rx_skb 14.77% [kernel] [k] napi_reuse_skb 8.95% [kernel] [k] skb_release_data 5.42% [kernel] [k] dev_gro_receive 5.37% [kernel] [k] memcpy <*> 5.26% [kernel] [k] iommu_dma_sync_sg_for_cpu 4.78% [kernel] [k] tcp_gro_receive <*> 4.42% [kernel] [k] iommu_dma_sync_sg_for_device 4.12% [kernel] [k] ipv6_gro_receive 3.65% [kernel] [k] gq_pool_get 3.25% [kernel] [k] skb_gro_receive 2.07% [kernel] [k] napi_gro_frags 1.98% [kernel] [k] tcp6_gro_receive 1.27% [kernel] [k] gq_rx_prep_buffers 1.18% [kernel] [k] gq_rx_napi_handler 0.99% [kernel] [k] csum_partial 0.74% [kernel] [k] csum_ipv6_magic 0.72% [kernel] [k] free_pcp_prepare 0.60% [kernel] [k] __napi_poll 0.58% [kernel] [k] net_rx_action 0.56% [kernel] [k] read_tsc <*> 0.50% [kernel] [k] __x86_indirect_thunk_r11 0.45% [kernel] [k] memset After patch, lines with <*> no longer show up, and overall cpu usage looks much better (~60% instead of ~72%) 25.56% [kernel] [k] gq_rx_skb 9.90% [kernel] [k] napi_reuse_skb 7.39% [kernel] [k] dev_gro_receive 6.78% [kernel] [k] memcpy 6.53% [kernel] [k] skb_release_data 6.39% [kernel] [k] tcp_gro_receive 5.71% [kernel] [k] ipv6_gro_receive 4.35% [kernel] [k] napi_gro_frags 4.34% [kernel] [k] skb_gro_receive 3.50% [kernel] [k] gq_pool_get 3.08% [kernel] [k] gq_rx_napi_handler 2.35% [kernel] [k] tcp6_gro_receive 2.06% [kernel] [k] gq_rx_prep_buffers 1.32% [kernel] [k] csum_partial 0.93% [kernel] [k] csum_ipv6_magic 0.65% [kernel] [k] net_rx_action On iavf, the UDP trafficgen with XDP_DROP in skb mode test shows +3-5% increase for direct DMA (when direct calls are avoided) and +10% with IOMMU (when indirect calls are avoided as well). Cc: Robin Murphy <robin.murphy@arm.com> Cc: Joerg Roedel <joro@8bytes.org> Cc: Will Deacon <will@kernel.org> Cc: iommu@lists.linux.dev Signed-off-by: Eric Dumazet <edumazet@google.com> Co-developed-by: Alexander Lobakin <aleksander.lobakin@intel.com> Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

page_pool::p is driver-defined params, copied directly from the structure passed to page_pool_create(). The structure isn't meant to be modified by the Page Pool core code and this even might look confusing[0][1]. In order to be able to alter some flags, let's define our own, internal fields the same way as the already existing one (::has_init_callback). They are defined as bits in the driver-set params, leave them so here as well, to not waste byte-per-bit or so. Almost 30 bits are still free for future extensions. We could've defined only new flags here or only the ones we may need to alter, but checking some flags in one place while others in another doesn't sound convenient or intuitive. ::flags passed by the driver can now go to the "slow" PP params. Suggested-by: Jakub Kicinski <kuba@kernel.org> Link[0]: https://lore.kernel.org/netdev/20230703133207.4f0c54ce@kernel.org Suggested-by: Alexander Duyck <alexanderduyck@fb.com> Link[1]: https://lore.kernel.org/netdev/CAKgT0UfZCGnWgOH96E4GV3ZP6LLbROHM7SHE8NKwq+exX+Gk_Q@mail.gmail.com Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Not a secret there's a ton of code duplication between two and more Intel ethernet modules. Before introducing new changes, which would need to be copied over again, start decoupling the already existing duplicate functionality into a new module, which will be shared between several Intel Ethernet drivers. Add the lookup table which converts 8/10-bit hardware packet type into a parsed bitfield structure for easy checking packet format parameters, such as payload level, IP version, etc. This is currently used by i40e, ice and iavf and it's all the same in all three drivers. The only difference introduced in this implementation is that instead of defining a 256 (or 1024 in case of ice) element array, add unlikely() condition to limit the input to 154 (current maximum non-reserved packet type). There's no reason to waste 600 (or even 3600) bytes only to not hurt very unlikely exception packets. The hash computation function now takes payload level directly as a pkt_hash_type. There's a couple cases when non-IP ptypes are marked as L3 payload and in the previous versions their hash level would be 2, not 3. But skb_set_hash() only sees difference between L4 and non-L4, thus this won't change anything at all. The module is behind the hidden Kconfig symbol, which the drivers will select when needed. The exports are behind 'LIBIE' namespace to limit the scope of the functions. Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Ever since build_skb() became stable, the old way with allocating an skb for storing the headers separately, which will be then copied manually, was slower, less flexible, and thus obsolete. * It had higher pressure on MM since it actually allocates new pages, which then get split and refcount-biased (NAPI page cache); * It implies memcpy() of packet headers (40+ bytes per each frame); * the actual header length was calculated via eth_get_headlen(), which invokes Flow Dissector and thus wastes a bunch of CPU cycles; * XDP makes it even more weird since it requires headroom for long and also tailroom for some time (since mbuf landed). Take a look at the ice driver, which is built around work-arounds to make XDP work with it. Even on some quite low-end hardware (not a common case for 100G NICs) it was performing worse. The only advantage "legacy-rx" had is that it didn't require any reserved headroom and tailroom. But iavf didn't use this, as it always splits pages into two halves of 2k, while that save would only be useful when striding. And again, XDP effectively removes that sole pro. There's a train of features to land in IAVF soon: Page Pool, XDP, XSk, multi-buffer etc. Each new would require adding more and more Danse Macabre for absolutely no reason, besides making hotpath less and less effective. Remove the "feature" with all the related code. This includes at least one very hot branch (typically hit on each new frame), which was either always-true or always-false at least for a complete NAPI bulk of 64 frames, the whole private flags cruft, and so on. Some stats: Function: add/remove: 0/4 grow/shrink: 0/7 up/down: 0/-721 (-721) RO Data: add/remove: 0/1 grow/shrink: 0/0 up/down: 0/-40 (-40) Reviewed-by: Alexander Duyck <alexanderduyck@fb.com> Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

As an intermediate step, remove all page splitting/recycling code. Just always allocate a new page and don't touch its refcount, so that it gets freed by the core stack later. Same for the "in-place" recycling, i.e. when an unused buffer gets assigned to a first needs-refilling descriptor. In some cases, this was leading to moving up to 63 &iavf_rx_buf structures around the ring on a per-field basis -- not something wanted on hotpath. The change allows to greatly simplify certain parts of the code: Function: add/remove: 0/2 grow/shrink: 0/7 up/down: 0/-744 (-744) Although the array of &iavf_rx_buf is barely used now and could be replaced with just page pointer array, don't touch it now to not complicate replacing it with libie Rx buffer struct later on. No surprise perf loses up to 30% here, but that regression will go away once PP lands. Note that iavf_rx_pg_*() definitions are left to reduce diffstat. They will be removed with the conversion to Page Pool. Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

There are several functions taking pointers to data they don't modify. This includes statistics fetching, page and page_pool parameters, etc. Constify the pointers, so that call sites will be able to pass const pointers as well. No functional changes, no visible changes in functions sizes. Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org> Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Each driver is responsible for syncing buffers written by HW for CPU before accessing them. Almost each PP-enabled driver uses the same pattern, which could be shorthanded into a static inline to make driver code a little bit more compact. Introduce a simple helper which performs DMA synchronization for the size passed from the driver. It can be used even when the pool doesn't manage DMA-syncs-for-device, just make sure the page has a correct DMA address set via page_pool_set_dma_addr(). Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Add a couple intuitive helpers to hide Rx buffer implementation details in the library and not multiplicate it between drivers. The settings are optimized for Intel hardware, but nothing really HW-specific here. Use the new page_pool_dev_alloc() to dynamically switch between split-page and full-page modes depending on MTU, page size, required headroom etc. For example, on x86_64 with the default driver settings each page is shared between 2 buffers. Turning on XDP (not in this series) -> increasing headroom requirement pushes truesize out of 2048 boundary, leading to that each buffer starts getting a full page. The "ceiling" limit is %PAGE_SIZE, as only order-0 pages are used to avoid compound overhead. For the above architecture, this means maximum linear frame size of 3712 w/o XDP. Not that &libie_buf_queue is not a complete queue/ring structure for now, rather a shim, but eventually the libie-enabled drivers will move to it, with iavf being the first one. Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Before replacing the Rx buffer management with libie, clean up &iavf_ring a bit. There are several fields not used anywhere in the code -- simply remove them. Move ::tail up to remove a hole. Replace ::arm_wb boolean with 1-bit flag in ::flags to free 1 more byte. Finally, move ::prev_pkt_ctr out of &iavf_tx_queue_stats -- it doesn't belong there (used for Tx stall detection). Place it next to the stats on the ring itself to fill the 4-byte slot. The result: no holes and all the hot fields fit into the first 64-byte cacheline. Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Now that the IAVF driver simply uses dev_alloc_page() + free_page() with no custom recycling logics, it can easily be switched to using Page Pool / libie API instead. This allows to removing the whole dancing around headroom, HW buffer size, and page order. All DMA-for-device is now done in the PP core, for-CPU -- in the libie helper. Use skb_mark_for_recycle() to bring back the recycling and restore the performance. Speaking of performance: on par with the baseline and faster with the PP optimization series applied. But the memory usage for 1500b MTU is now almost 2x lower (x86_64) thanks to allocating a page every second descriptor. Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Next stop, per-queue private stats. They have only subtle differences from driver to driver and can easily be resolved. Define common structures, inline helpers and Ethtool helpers to collect, update and export the statistics. Use u64_stats_t right from the start, as well as the corresponding helpers to ensure tear-free operations. For the NAPI parts of both Rx and Tx, also define small onstack containers to update them in polling loops and then sync the actual containers once a loop ends. The drivers will be switched to use this API later on a per-driver basis, along with conversion to PP. Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

iavf is pretty much ready for using the generic libie stats, so drop all the custom code and just use generic definitions. The only thing is that it previously lacked the counter of Tx queue stops. It's present in the other drivers, so add it here as well. The rest is straightforward. Note that it makes the ring structure 1 CL bigger, but no layout changes since the stats structures are 32-byte aligned. Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Tell hardware to write back completed descriptors even when interrupts are disabled. Otherwise, descriptors might not be written back until the hardware can flush a full cacheline of descriptors. This can cause unnecessary delays when traffic is light (or even trigger Tx queue timeout). The example scenario to reproduce the Tx timeout if the fix is not applied: - configure at least 2 Tx queues to be assigned to the same q_vector, - generate a huge Tx traffic on the first Tx queue - try to send a few packets using the second Tx queue. In such a case Tx timeout will appear on the second Tx queue because no completion descriptors are written back for that queue while interrupts are disabled due to NAPI polling. The patch is necessary to start work on the AF_XDP implementation for the idpf driver, because there may be a case where a regular LAN Tx queue and an XDP queue share the same NAPI. Fixes: c2d548c ("idpf: add TX splitq napi poll support") Fixes: a5ab9ee ("idpf: add singleq start_xmit and napi poll") Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com> Signed-off-by: Joshua Hay <joshua.a.hay@intel.com> Co-developed-by: Michal Kubiak <michal.kubiak@intel.com> Signed-off-by: Michal Kubiak <michal.kubiak@intel.com>

ui_browser__show() is capturing the input title that is stack allocated memory in hist_browser__run(). Avoid a use after return by strdup-ing the string. Committer notes: Further explanation from Ian Rogers: My command line using tui is: $ sudo bash -c 'rm /tmp/asan.log*; export ASAN_OPTIONS="log_path=/tmp/asan.log"; /tmp/perf/perf mem record -a sleep 1; /tmp/perf/perf mem report' I then go to the perf annotate view and quit. This triggers the asan error (from the log file): ``` ==1254591==ERROR: AddressSanitizer: stack-use-after-return on address 0x7f2813331920 at pc 0x7f28180 65991 bp 0x7fff0a21c750 sp 0x7fff0a21bf10 READ of size 80 at 0x7f2813331920 thread T0 #0 0x7f2818065990 in __interceptor_strlen ../../../../src/libsanitizer/sanitizer_common/sanitizer_common_interceptors.inc:461 #1 0x7f2817698251 in SLsmg_write_wrapped_string (/lib/x86_64-linux-gnu/libslang.so.2+0x98251) #2 0x7f28176984b9 in SLsmg_write_nstring (/lib/x86_64-linux-gnu/libslang.so.2+0x984b9) #3 0x55c94045b365 in ui_browser__write_nstring ui/browser.c:60 #4 0x55c94045c558 in __ui_browser__show_title ui/browser.c:266 #5 0x55c94045c776 in ui_browser__show ui/browser.c:288 #6 0x55c94045c06d in ui_browser__handle_resize ui/browser.c:206 #7 0x55c94047979b in do_annotate ui/browsers/hists.c:2458 #8 0x55c94047fb17 in evsel__hists_browse ui/browsers/hists.c:3412 #9 0x55c940480a0c in perf_evsel_menu__run ui/browsers/hists.c:3527 #10 0x55c940481108 in __evlist__tui_browse_hists ui/browsers/hists.c:3613 #11 0x55c9404813f7 in evlist__tui_browse_hists ui/browsers/hists.c:3661 #12 0x55c93ffa253f in report__browse_hists tools/perf/builtin-report.c:671 #13 0x55c93ffa58ca in __cmd_report tools/perf/builtin-report.c:1141 #14 0x55c93ffaf159 in cmd_report tools/perf/builtin-report.c:1805 #15 0x55c94000c05c in report_events tools/perf/builtin-mem.c:374 #16 0x55c94000d96d in cmd_mem tools/perf/builtin-mem.c:516 #17 0x55c9400e44ee in run_builtin tools/perf/perf.c:350 #18 0x55c9400e4a5a in handle_internal_command tools/perf/perf.c:403 #19 0x55c9400e4e22 in run_argv tools/perf/perf.c:447 #20 0x55c9400e53ad in main tools/perf/perf.c:561 #21 0x7f28170456c9 in __libc_start_call_main ../sysdeps/nptl/libc_start_call_main.h:58 #22 0x7f2817045784 in __libc_start_main_impl ../csu/libc-start.c:360 #23 0x55c93ff544c0 in _start (/tmp/perf/perf+0x19a4c0) (BuildId: 84899b0e8c7d3a3eaa67b2eb35e3d8b2f8cd4c93) Address 0x7f2813331920 is located in stack of thread T0 at offset 32 in frame #0 0x55c94046e85e in hist_browser__run ui/browsers/hists.c:746 This frame has 1 object(s): [32, 192) 'title' (line 747) <== Memory access at offset 32 is inside this variable HINT: this may be a false positive if your program uses some custom stack unwind mechanism, swapcontext or vfork ``` hist_browser__run isn't on the stack so the asan error looks legit. There's no clean init/exit on struct ui_browser so I may be trading a use-after-return for a memory leak, but that seems look a good trade anyway. Fixes: 05e8b08 ("perf ui browser: Stop using 'self'") Signed-off-by: Ian Rogers <irogers@google.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Athira Rajeev <atrajeev@linux.vnet.ibm.com> Cc: Ben Gainey <ben.gainey@arm.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: James Clark <james.clark@arm.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Kajol Jain <kjain@linux.ibm.com> Cc: Kan Liang <kan.liang@linux.intel.com> Cc: K Prateek Nayak <kprateek.nayak@amd.com> Cc: Li Dong <lidong@vivo.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Oliver Upton <oliver.upton@linux.dev> Cc: Paran Lee <p4ranlee@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ravi Bangoria <ravi.bangoria@amd.com> Cc: Sun Haiyong <sunhaiyong@loongson.cn> Cc: Tim Chen <tim.c.chen@linux.intel.com> Cc: Yanteng Si <siyanteng@loongson.cn> Cc: Yicong Yang <yangyicong@hisilicon.com> Link: https://lore.kernel.org/r/20240507183545.1236093-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

With latest llvm19, the selftest iters/iter_arr_with_actual_elem_count failed with -mcpu=v4. The following are the details: 0: R1=ctx() R10=fp0 ; int iter_arr_with_actual_elem_count(const void *ctx) @ iters.c:1420 0: (b4) w7 = 0 ; R7_w=0 ; int i, n = loop_data.n, sum = 0; @ iters.c:1422 1: (18) r1 = 0xffffc90000191478 ; R1_w=map_value(map=iters.bss,ks=4,vs=1280,off=1144) 3: (61) r6 = *(u32 *)(r1 +128) ; R1_w=map_value(map=iters.bss,ks=4,vs=1280,off=1144) R6_w=scalar(smin=0,smax=umax=0xffffffff,var_off=(0x0; 0xffffffff)) ; if (n > ARRAY_SIZE(loop_data.data)) @ iters.c:1424 4: (26) if w6 > 0x20 goto pc+27 ; R6_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=32,var_off=(0x0; 0x3f)) 5: (bf) r8 = r10 ; R8_w=fp0 R10=fp0 6: (07) r8 += -8 ; R8_w=fp-8 ; bpf_for(i, 0, n) { @ iters.c:1427 7: (bf) r1 = r8 ; R1_w=fp-8 R8_w=fp-8 8: (b4) w2 = 0 ; R2_w=0 9: (bc) w3 = w6 ; R3_w=scalar(id=1,smin=smin32=0,smax=umax=smax32=umax32=32,var_off=(0x0; 0x3f)) R6_w=scalar(id=1,smin=smin32=0,smax=umax=smax32=umax32=32,var_off=(0x0; 0x3f)) 10: (85) call bpf_iter_num_new#45179 ; R0=scalar() fp-8=iter_num(ref_id=2,state=active,depth=0) refs=2 11: (bf) r1 = r8 ; R1=fp-8 R8=fp-8 refs=2 12: (85) call bpf_iter_num_next#45181 13: R0=rdonly_mem(id=3,ref_obj_id=2,sz=4) R6=scalar(id=1,smin=smin32=0,smax=umax=smax32=umax32=32,var_off=(0x0; 0x3f)) R7=0 R8=fp-8 R10=fp0 fp-8=iter_num(ref_id=2,state=active,depth=1) refs=2 ; bpf_for(i, 0, n) { @ iters.c:1427 13: (15) if r0 == 0x0 goto pc+2 ; R0=rdonly_mem(id=3,ref_obj_id=2,sz=4) refs=2 14: (81) r1 = *(s32 *)(r0 +0) ; R0=rdonly_mem(id=3,ref_obj_id=2,sz=4) R1_w=scalar(smin=0xffffffff80000000,smax=0x7fffffff) refs=2 15: (ae) if w1 < w6 goto pc+4 20: R0=rdonly_mem(id=3,ref_obj_id=2,sz=4) R1=scalar(smin=0xffffffff80000000,smax=smax32=umax32=31,umax=0xffffffff0000001f,smin32=0,var_off=(0x0; 0xffffffff0000001f)) R6=scalar(id=1,smin=umin=smin32=umin32=1,smax=umax=smax32=umax32=32,var_off=(0x0; 0x3f)) R7=0 R8=fp-8 R10=fp0 fp-8=iter_num(ref_id=2,state=active,depth=1) refs=2 ; sum += loop_data.data[i]; @ iters.c:1429 20: (67) r1 <<= 2 ; R1_w=scalar(smax=0x7ffffffc0000007c,umax=0xfffffffc0000007c,smin32=0,smax32=umax32=124,var_off=(0x0; 0xfffffffc0000007c)) refs=2 21: (18) r2 = 0xffffc90000191478 ; R2_w=map_value(map=iters.bss,ks=4,vs=1280,off=1144) refs=2 23: (0f) r2 += r1 math between map_value pointer and register with unbounded min value is not allowed The source code: int iter_arr_with_actual_elem_count(const void *ctx) { int i, n = loop_data.n, sum = 0; if (n > ARRAY_SIZE(loop_data.data)) return 0; bpf_for(i, 0, n) { /* no rechecking of i against ARRAY_SIZE(loop_data.n) */ sum += loop_data.data[i]; } return sum; } The insn #14 is a sign-extenstion load which is related to 'int i'. The insn #15 did a subreg comparision. Note that smin=0xffffffff80000000 and this caused later insn #23 failed verification due to unbounded min value. Actually insn #15 R1 smin range can be better. Before insn #15, we have R1_w=scalar(smin=0xffffffff80000000,smax=0x7fffffff) With the above range, we know for R1, upper 32bit can only be 0xffffffff or 0. Otherwise, the value range for R1 could be beyond [smin=0xffffffff80000000,smax=0x7fffffff]. After insn #15, for the true patch, we know smin32=0 and smax32=32. With the upper 32bit 0xffffffff, then the corresponding value is [0xffffffff00000000, 0xffffffff00000020]. The range is obviously beyond the original range [smin=0xffffffff80000000,smax=0x7fffffff] and the range is not possible. So the upper 32bit must be 0, which implies smin = smin32 and smax = smax32. This patch fixed the issue by adding additional register deduction after 32-bit compare insn. If the signed 32-bit register range is non-negative then 64-bit smin is in range of [S32_MIN, S32_MAX], then the actual 64-bit smin/smax should be the same as 32-bit smin32/smax32. With this patch, iters/iter_arr_with_actual_elem_count succeeded with better register range: from 15 to 20: R0=rdonly_mem(id=7,ref_obj_id=2,sz=4) R1_w=scalar(smin=smin32=0,smax=umax=smax32=umax32=31,var_off=(0x0; 0x1f)) R6=scalar(id=1,smin=umin=smin32=umin32=1,smax=umax=smax32=umax32=32,var_off=(0x0; 0x3f)) R7=scalar(id=9,smin=0,smax=umax=0xffffffff,var_off=(0x0; 0xffffffff)) R8=scalar(id=9,smin=0,smax=umax=0xffffffff,var_off=(0x0; 0xffffffff)) R10=fp0 fp-8=iter_num(ref_id=2,state=active,depth=3) refs=2 Acked-by: Eduard Zingerman <eddyz87@gmail.com> Acked-by: Shung-Hsi Yu <shung-hsi.yu@suse.com> Signed-off-by: Yonghong Song <yonghong.song@linux.dev> Link: https://lore.kernel.org/r/20240723162933.2731620-1-yonghong.song@linux.dev Signed-off-by: Alexei Starovoitov <ast@kernel.org>

alobakin and others added 19 commits January 12, 2024 13:03

dma: compile-out DMA sync operations when not used

9a3aa94

Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

page_pool: check for DMA sync shortcut earlier

78ad4db

Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

xsk: use generic DMA sync shortcut instead of a custom one

af9cc6e

Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>

Port Josh'es refactoring

afa0d4f

Port Josh's fix for tx_buf refactor

a24cee9

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Wb on itr experiments #23

Wb on itr experiments #23

michalQb commented Feb 20, 2024

Wb on itr experiments #23

Are you sure you want to change the base?

Wb on itr experiments #23

Conversation

michalQb commented Feb 20, 2024