1 .. SPDX-License-Identifier: GPL-2.0 2 .. include:: <isonum.txt> 3 4 ================================== 5 CXL Access Coordinates Computation 6 ================================== 7 8 Shared Upstream Link Calculation 9 ================================ 10 For certain CXL region construction with endpoints behind CXL switches (SW) or 11 Root Ports (RP), there is the possibility of the total bandwidth for all 12 the endpoints behind a switch being more than the switch upstream link. 13 A similar situation can occur within the host, upstream of the root ports. 14 The CXL driver performs an additional pass after all the targets have 15 arrived for a region in order to recalculate the bandwidths with possible 16 upstream link being a limiting factor in mind. 17 18 The algorithm assumes the configuration is a symmetric topology as that 19 maximizes performance. When asymmetric topology is detected, the calculation 20 is aborted. An asymmetric topology is detected during topology walk where the 21 number of RPs detected as a grandparent is not equal to the number of devices 22 iterated in the same iteration loop. The assumption is made that subtle 23 asymmetry in properties does not happen and all paths to EPs are equal. 24 25 There can be multiple switches under an RP. There can be multiple RPs under 26 a CXL Host Bridge (HB). There can be multiple HBs under a CXL Fixed Memory 27 Window Structure (CFMWS). 28 29 An example hierarchy: 30 31 > CFMWS 0 32 > | 33 > _________|_________ 34 > | | 35 > ACPI0017-0 ACPI0017-1 36 > GP0/HB0/ACPI0016-0 GP1/HB1/ACPI0016-1 37 > | | | | 38 > RP0 RP1 RP2 RP3 39 > | | | | 40 > SW 0 SW 1 SW 2 SW 3 41 > | | | | | | | | 42 > EP0 EP1 EP2 EP3 EP4 EP5 EP6 EP7 43 44 Computation for the example hierarchy: 45 46 Min (GP0 to CPU BW, 47 Min(SW 0 Upstream Link to RP0 BW, 48 Min(SW0SSLBIS for SW0DSP0 (EP0), EP0 DSLBIS, EP0 Upstream Link) + 49 Min(SW0SSLBIS for SW0DSP1 (EP1), EP1 DSLBIS, EP1 Upstream link)) + 50 Min(SW 1 Upstream Link to RP1 BW, 51 Min(SW1SSLBIS for SW1DSP0 (EP2), EP2 DSLBIS, EP2 Upstream Link) + 52 Min(SW1SSLBIS for SW1DSP1 (EP3), EP3 DSLBIS, EP3 Upstream link))) + 53 Min (GP1 to CPU BW, 54 Min(SW 2 Upstream Link to RP2 BW, 55 Min(SW2SSLBIS for SW2DSP0 (EP4), EP4 DSLBIS, EP4 Upstream Link) + 56 Min(SW2SSLBIS for SW2DSP1 (EP5), EP5 DSLBIS, EP5 Upstream link)) + 57 Min(SW 3 Upstream Link to RP3 BW, 58 Min(SW3SSLBIS for SW3DSP0 (EP6), EP6 DSLBIS, EP6 Upstream Link) + 59 Min(SW3SSLBIS for SW3DSP1 (EP7), EP7 DSLBIS, EP7 Upstream link)))) 60 61 The calculation starts at cxl_region_shared_upstream_perf_update(). A xarray 62 is created to collect all the endpoint bandwidths via the 63 cxl_endpoint_gather_bandwidth() function. The min() of bandwidth from the 64 endpoint CDAT and the upstream link bandwidth is calculated. If the endpoint 65 has a CXL switch as a parent, then min() of calculated bandwidth and the 66 bandwidth from the SSLBIS for the switch downstream port that is associated 67 with the endpoint is calculated. The final bandwidth is stored in a 68 'struct cxl_perf_ctx' in the xarray indexed by a device pointer. If the 69 endpoint is direct attached to a root port (RP), the device pointer would be an 70 RP device. If the endpoint is behind a switch, the device pointer would be the 71 upstream device of the parent switch. 72 73 At the next stage, the code walks through one or more switches if they exist 74 in the topology. For endpoints directly attached to RPs, this step is skipped. 75 If there is another switch upstream, the code takes the min() of the current 76 gathered bandwidth and the upstream link bandwidth. If there's a switch 77 upstream, then the SSLBIS of the upstream switch. 78 79 Once the topology walk reaches the RP, whether it's direct attached endpoints 80 or walking through the switch(es), cxl_rp_gather_bandwidth() is called. At 81 this point all the bandwidths are aggregated per each host bridge, which is 82 also the index for the resulting xarray. 83 84 The next step is to take the min() of the per host bridge bandwidth and the 85 bandwidth from the Generic Port (GP). The bandwidths for the GP is retrieved 86 via ACPI tables SRAT/HMAT. The min bandwidth are aggregated under the same 87 ACPI0017 device to form a new xarray. 88 89 Finally, the cxl_region_update_bandwidth() is called and the aggregated 90 bandwidth from all the members of the last xarray is updated for the 91 access coordinates residing in the cxl region (cxlr) context.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.