| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <43cb44d6-86f2-66a3-0167-6c75a97b3c9a@amd.com> Date: Tue, 29 Oct 2024 08:24:37 -0700 From: Lizhi Hou <lizhi.hou@....com> To: Jeffrey Hugo <quic_jhugo@...cinc.com>, <ogabbay@...nel.org>, <dri-devel@...ts.freedesktop.org> CC: <linux-kernel@...r.kernel.org>, <min.ma@....com>, <max.zhen@....com>, <sonal.santan@....com>, <king.tam@....com> Subject: Re: [PATCH V5 00/10] AMD XDNA driver On 10/25/24 15:02, Jeffrey Hugo wrote: > On 10/25/2024 3:28 PM, Lizhi Hou wrote: >> >> On 10/25/24 10:55, Jeffrey Hugo wrote: >>> On 10/21/2024 10:19 AM, Lizhi Hou wrote: >>>> This patchset introduces a new Linux Kernel Driver, amdxdna for AMD >>>> NPUs. >>>> The driver is based on Linux accel subsystem. >>>> >>>> NPU (Neural Processing Unit) is an AI inference accelerator integrated >>>> into AMD client CPUs. NPU enables efficient execution of Machine >>>> Learning >>>> applications like CNNs, LLMs, etc. NPU is based on AMD XDNA >>>> architecture [1]. >>>> >>>> AMD NPU consists of the following components: >>>> >>>> - Tiled array of AMD AI Engine processors. >>>> - Micro Controller which runs the NPU Firmware responsible for >>>> command processing, AIE array configuration, and execution >>>> management. >>>> - PCI EP for host control of the NPU device. >>>> - Interconnect for connecting the NPU components together. >>>> - SRAM for use by the NPU Firmware. >>>> - Address translation hardware for protected host memory access >>>> by the >>>> NPU. >>>> >>>> NPU supports multiple concurrent fully isolated contexts. Concurrent >>>> contexts may be bound to AI Engine array spatially and or temporarily. >>>> >>>> The driver is licensed under GPL-2.0 except for UAPI header which is >>>> licensed GPL-2.0 WITH Linux-syscall-note. >>>> >>>> User mode driver stack consists of XRT [2] and AMD AIE Plugin for >>>> IREE [3]. >>>> >>>> The firmware for the NPU is distributed as a closed source binary, >>>> and has >>>> already been pushed to the DRM firmware repository [4]. >>>> >>>> [1]https://www.amd.com/en/technologies/xdna.html >>>> [2]https://github.com/Xilinx/XRT >>>> [3]https://github.com/nod-ai/iree-amd-aie >>>> [4]https://gitlab.freedesktop.org/drm/firmware/-/tree/amd-ipu-staging/amdnpu >>>> >>>> >>>> Changes since v4: >>>> - Fix lockdep errors >>>> - Use __u* structure for struct aie_error >>> >>> One nit, when you send the next version would you please either To: >>> or Cc: me on the entire series? I only get pieces in my inbox which >>> is mildly annoying on my end. >> Sure. >>> >>> Looks like we are getting close here. One procedural question I >>> have, do you have commit permissions to drm-misc? >> No, I do not have commit permissions yet. > > You should apply for access. Assuming this series is ready before > that goes through, I'll apply it. > >>> I applied the series to drm-misc-next and tried to build. Got the >>> following errors - >> >> Could you share the build command line? So I can reproduce and verify >> my fix. > > The command is simple: > make -j20 > > The system details, incase it somehow matters: > Ubuntu 22.04 w/ 5.15 kernel > > $ lsb_release -a > No LSB modules are available. > Distributor ID: Ubuntu > Description: Ubuntu 22.04.3 LTS > Release: 22.04 > Codename: jammy > > $ uname -a > Linux jhugo-lnx 5.15.0-89-generic #99-Ubuntu SMP Mon Oct 30 20:42:41 > UTC 2023 x86_64 x86_64 x86_64 GNU/Linux > > The kernel config is probably the relevant piece. When I first built > after applying the series, I was asked to choose what to do with the > new config item. I selected =m. > .config can be found at > https://gist.github.com/quic-jhugo/4cc249b1e3ba127039fbc709a513a432 > >> >> I used "make M=drivers/accel/amdxdna" and did not reproduce the error >> with drm-misc-next. It looks build robot did not complain with the >> patch neither. >> >> $ git branch >> * drm-misc-next >> $ make M=drivers/accel/amdxdna >> CC [M] drivers/accel/amdxdna/aie2_ctx.o >> CC [M] drivers/accel/amdxdna/aie2_error.o >> CC [M] drivers/accel/amdxdna/aie2_message.o >> CC [M] drivers/accel/amdxdna/aie2_pci.o >> CC [M] drivers/accel/amdxdna/aie2_psp.o >> CC [M] drivers/accel/amdxdna/aie2_smu.o >> CC [M] drivers/accel/amdxdna/aie2_solver.o >> CC [M] drivers/accel/amdxdna/amdxdna_ctx.o >> CC [M] drivers/accel/amdxdna/amdxdna_gem.o >> CC [M] drivers/accel/amdxdna/amdxdna_mailbox.o >> CC [M] drivers/accel/amdxdna/amdxdna_mailbox_helper.o >> CC [M] drivers/accel/amdxdna/amdxdna_pci_drv.o >> CC [M] drivers/accel/amdxdna/amdxdna_sysfs.o >> CC [M] drivers/accel/amdxdna/npu1_regs.o >> CC [M] drivers/accel/amdxdna/npu2_regs.o >> CC [M] drivers/accel/amdxdna/npu4_regs.o >> CC [M] drivers/accel/amdxdna/npu5_regs.o >> LD [M] drivers/accel/amdxdna/amdxdna.o >> MODPOST drivers/accel/amdxdna/Module.symvers >> CC [M] drivers/accel/amdxdna/amdxdna.mod.o >> CC [M] drivers/accel/amdxdna/.module-common.o >> LD [M] drivers/accel/amdxdna/amdxdna.ko >> $ >> >>> >>> CC [M] drivers/accel/amdxdna/aie2_ctx.o >>> CC [M] drivers/accel/amdxdna/aie2_error.o >>> CC [M] drivers/accel/amdxdna/aie2_message.o >>> CC [M] drivers/accel/amdxdna/aie2_pci.o >>> CC [M] drivers/accel/amdxdna/aie2_psp.o >>> CC [M] drivers/accel/amdxdna/aie2_smu.o >>> CC [M] drivers/accel/amdxdna/aie2_solver.o >>> CC [M] drivers/accel/amdxdna/amdxdna_ctx.o >>> CC [M] drivers/accel/amdxdna/amdxdna_gem.o >>> CC [M] drivers/accel/amdxdna/amdxdna_mailbox.o >>> CC [M] drivers/accel/amdxdna/amdxdna_mailbox_helper.o >>> CC [M] drivers/accel/amdxdna/amdxdna_pci_drv.o >>> CC [M] drivers/accel/amdxdna/amdxdna_sysfs.o >>> CC [M] drivers/accel/amdxdna/npu1_regs.o >>> CC [M] drivers/accel/amdxdna/npu2_regs.o >>> CC [M] drivers/accel/amdxdna/npu4_regs.o >>> CC [M] drivers/accel/amdxdna/npu5_regs.o >>> AR drivers/base/firmware_loader/built-in.a >>> AR drivers/base/built-in.a >>> In file included from drivers/accel/amdxdna/aie2_message.c:19: >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function ‘amdxdna_cmd_get_op’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:112:16: error: implicit >>> declaration of function ‘FIELD_GET’ >>> [-Werror=implicit-function-declaration] >>> 112 | return FIELD_GET(AMDXDNA_CMD_OPCODE, cmd->header); >>> | ^~~~~~~~~ >>> In file included from drivers/accel/amdxdna/amdxdna_gem.c:15: >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function ‘amdxdna_cmd_get_op’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:112:16: error: implicit >>> declaration of function ‘FIELD_GET’ >>> [-Werror=implicit-function-declaration] >>> 112 | return FIELD_GET(AMDXDNA_CMD_OPCODE, cmd->header); >>> | ^~~~~~~~~ >>> In file included from drivers/accel/amdxdna/aie2_psp.c:11: >>> drivers/accel/amdxdna/aie2_psp.c: In function ‘psp_exec’: >>> drivers/accel/amdxdna/aie2_psp.c:62:34: error: implicit declaration >>> of function ‘FIELD_GET’ [-Werror=implicit-function-declaration] >>> 62 | FIELD_GET(PSP_STATUS_READY, ready), >>> | ^~~~~~~~~ >>> ./include/linux/iopoll.h:47:21: note: in definition of macro >>> ‘read_poll_timeout’ >>> 47 | if (cond) \ >>> | ^~~~ >>> drivers/accel/amdxdna/aie2_psp.c:61:15: note: in expansion of macro >>> ‘readx_poll_timeout’ >>> 61 | ret = readx_poll_timeout(readl, PSP_REG(psp, >>> PSP_STATUS_REG), ready, >>> | ^~~~~~~~~~~~~~~~~~ >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function >>> ‘amdxdna_cmd_set_state’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:121:24: error: implicit >>> declaration of function ‘FIELD_PREP’ >>> [-Werror=implicit-function-declaration] >>> 121 | cmd->header |= FIELD_PREP(AMDXDNA_CMD_STATE, s); >>> | ^~~~~~~~~~ >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function >>> ‘amdxdna_cmd_set_state’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:121:24: error: implicit >>> declaration of function ‘FIELD_PREP’ >>> [-Werror=implicit-function-declaration] >>> 121 | cmd->header |= FIELD_PREP(AMDXDNA_CMD_STATE, s); >>> | ^~~~~~~~~~ >>> In file included from drivers/accel/amdxdna/aie2_pci.c:22: >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function ‘amdxdna_cmd_get_op’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:112:16: error: implicit >>> declaration of function ‘FIELD_GET’ >>> [-Werror=implicit-function-declaration] >>> 112 | return FIELD_GET(AMDXDNA_CMD_OPCODE, cmd->header); >>> | ^~~~~~~~~ >>> In file included from drivers/accel/amdxdna/aie2_ctx.c:18: >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function ‘amdxdna_cmd_get_op’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:112:16: error: implicit >>> declaration of function ‘FIELD_GET’ >>> [-Werror=implicit-function-declaration] >>> 112 | return FIELD_GET(AMDXDNA_CMD_OPCODE, cmd->header); >>> | ^~~~~~~~~ >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function >>> ‘amdxdna_cmd_set_state’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:121:24: error: implicit >>> declaration of function ‘FIELD_PREP’ >>> [-Werror=implicit-function-declaration] >>> 121 | cmd->header |= FIELD_PREP(AMDXDNA_CMD_STATE, s); >>> | ^~~~~~~~~~ >>> In file included from drivers/accel/amdxdna/amdxdna_ctx.c:16: >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function ‘amdxdna_cmd_get_op’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:112:16: error: implicit >>> declaration of function ‘FIELD_GET’ >>> [-Werror=implicit-function-declaration] >>> 112 | return FIELD_GET(AMDXDNA_CMD_OPCODE, cmd->header); >>> | ^~~~~~~~~ >>> cc1: all warnings being treated as errors >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function >>> ‘amdxdna_cmd_set_state’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:121:24: error: implicit >>> declaration of function ‘FIELD_PREP’ >>> [-Werror=implicit-function-declaration] >>> 121 | cmd->header |= FIELD_PREP(AMDXDNA_CMD_STATE, s); >>> | ^~~~~~~~~~ >>> drivers/accel/amdxdna/aie2_ctx.c: In function ‘aie2_hwctx_restart’: >>> drivers/accel/amdxdna/aie2_ctx.c:114:9: error: too few arguments to >>> function ‘drm_sched_start’ >>> 114 | drm_sched_start(&hwctx->priv->sched); >>> | ^~~~~~~~~~~~~~~ >>> In file included from ./include/trace/events/amdxdna.h:12, >>> from drivers/accel/amdxdna/aie2_ctx.c:13: >>> ./include/drm/gpu_scheduler.h:593:6: note: declared here >>> 593 | void drm_sched_start(struct drm_gpu_scheduler *sched, int >>> errno); >>> | ^~~~~~~~~~~~~~~ >>> make[5]: *** [scripts/Makefile.build:229: >>> drivers/accel/amdxdna/aie2_psp.o] Error 1 >>> make[5]: *** Waiting for unfinished jobs.... >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function >>> ‘amdxdna_cmd_set_state’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:121:24: error: implicit >>> declaration of function ‘FIELD_PREP’ >>> [-Werror=implicit-function-declaration] >>> 121 | cmd->header |= FIELD_PREP(AMDXDNA_CMD_STATE, s); >>> | ^~~~~~~~~~ >>> In file included from drivers/accel/amdxdna/amdxdna_pci_drv.c:18: >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function ‘amdxdna_cmd_get_op’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:112:16: error: implicit >>> declaration of function ‘FIELD_GET’ >>> [-Werror=implicit-function-declaration] >>> 112 | return FIELD_GET(AMDXDNA_CMD_OPCODE, cmd->header); >>> | ^~~~~~~~~ >>> cc1: all warnings being treated as errors >>> make[5]: *** [scripts/Makefile.build:229: >>> drivers/accel/amdxdna/aie2_ctx.o] Error 1 >>> drivers/accel/amdxdna/amdxdna_ctx.h: In function >>> ‘amdxdna_cmd_set_state’: >>> drivers/accel/amdxdna/amdxdna_ctx.h:121:24: error: implicit >>> declaration of function ‘FIELD_PREP’ >>> [-Werror=implicit-function-declaration] >>> 121 | cmd->header |= FIELD_PREP(AMDXDNA_CMD_STATE, s); >>> | ^~~~~~~~~~ >>> drivers/accel/amdxdna/amdxdna_mailbox.c: In function >>> ‘xdna_mailbox_send_msg’: >>> drivers/accel/amdxdna/amdxdna_mailbox.c:444:26: error: implicit >>> declaration of function ‘FIELD_PREP’ >>> [-Werror=implicit-function-declaration] >>> 444 | header->sz_ver = FIELD_PREP(MSG_BODY_SZ, >>> msg->send_size) | >>> | ^~~~~~~~~~ >>> >>> >>> You also have the following checkpatch issues - >> >> Could you share the command you used? I tried to use 'dim >> checkpatch' and it did not find out the misspelling issue. > > ./scripts/checkpatch.pl --strict --codespell *.patch > > Note, --codespell requires some local setup. I beleive the comments > in the checkpatch.pl script are fairly straightforward. I use a copy > of the database from the github that is rather recent. The Ubuntu > distro package is really out of date and I don't think I looked to see > if there is a pythong pip version. Grabbing the one file from the > github repo seemed simple emough. I was able to reproduce with your suggestions. Thanks a lot. Lizhi > > -Jeff
Powered by blists - more mailing lists