From nobody Thu Nov 28 02:54:03 2024 Received: from mx0a-0031df01.pphosted.com (mx0a-0031df01.pphosted.com [205.220.168.131]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id CB385187353; Mon, 7 Oct 2024 08:45:42 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=205.220.168.131 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1728290744; cv=none; b=f+RejDWTBKZy43JTw2Uuv0BA9ZxiJt0jXENzji8Tt2vICMT3Yyt+rDDds52HCaP+l9msVsGihYs0PLveIoV8lcJrR+jdpUh9ZACKrF1pjjtmrxNS/5ABAuCVCBHldM9BKEyaRdsBehtSfZOGC/qGxIPfOAU74Fn0ZM0qpBLjTTo= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1728290744; c=relaxed/simple; bh=4He6GAK85rGaihAzpW1qOad9XowHPfchucXUb9v6EyY=; h=From:To:CC:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version:Content-Type; b=dIrmBe3a2Hxwo+bCtN7UG7GIqaxQUOe8SBWyPS76/xe8RMXD26/GtBOXPFp0R3QNDXuzMdU4CctBVq+4o8in88uR853LmD55W5rDIYEqT6msCgoNpcDXp4x3icDE7GwLL9jh7kEVCz1FPFvTVmZ2E2ZbrGbpvJTAvhNNQzpfMqE= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=quicinc.com; spf=pass smtp.mailfrom=quicinc.com; dkim=pass (2048-bit key) header.d=quicinc.com header.i=@quicinc.com header.b=WnIWLH/B; arc=none smtp.client-ip=205.220.168.131 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=quicinc.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=quicinc.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=quicinc.com header.i=@quicinc.com header.b="WnIWLH/B" Received: from pps.filterd (m0279867.ppops.net [127.0.0.1]) by mx0a-0031df01.pphosted.com (8.18.1.2/8.18.1.2) with ESMTP id 4974Sr8g032224; Mon, 7 Oct 2024 08:45:36 GMT DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=quicinc.com; h= cc:content-transfer-encoding:content-type:date:from:in-reply-to :message-id:mime-version:references:subject:to; s=qcppdkim1; bh= d/xYOlAfHq811yxJ/AL8jc7fZL4S8qi38cwZvN2sd8c=; b=WnIWLH/BsLAF0/gm FJWLPObfKoXjPVuPwBqLK35CmESDlIhc95B4p4EiVlneNzcEIm41eOBAndWBVqlL mIVivOwxfbWuw+Qy4T00FfMg3nYg55CqKuf7Tm5nZGlBWP29LfepvkIzgY92ZHQx 8grAY7sq9Lc69TrKTgwcr++Z9XMemahsgRZXBN1s8LyUsEWf1Xkvqg4A9Y682LPh SXhc13F+/Tgb35VRX7zJr/0Fk1vaN8bTJnWOKnmS3rCTg/yoi4PGZ9ItzlDUOFoP ZIZQueXxz2RdMWzGEoXBwfqFyg5I15K3O98y9gzx0wzVwMCMQ4CNZRzVUj1CMusl rUJm2Q== Received: from nalasppmta03.qualcomm.com (Global_NAT1.qualcomm.com [129.46.96.20]) by mx0a-0031df01.pphosted.com (PPS) with ESMTPS id 422xv8bff5-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 07 Oct 2024 08:45:36 +0000 (GMT) Received: from nalasex01b.na.qualcomm.com (nalasex01b.na.qualcomm.com [10.47.209.197]) by NALASPPMTA03.qualcomm.com (8.18.1.2/8.18.1.2) with ESMTPS id 4978jZwd024819 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 7 Oct 2024 08:45:35 GMT Received: from hu-ekangupt-hyd.qualcomm.com (10.80.80.8) by nalasex01b.na.qualcomm.com (10.47.209.197) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.2.1544.9; Mon, 7 Oct 2024 01:45:33 -0700 From: Ekansh Gupta To: , CC: , , , , , Subject: [PATCH v1 2/4] misc: fastrpc: Capture kernel and DSP performance counters Date: Mon, 7 Oct 2024 14:15:16 +0530 Message-ID: <20241007084518.3649876-3-quic_ekangupt@quicinc.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20241007084518.3649876-1-quic_ekangupt@quicinc.com> References: <20241007084518.3649876-1-quic_ekangupt@quicinc.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable X-ClientProxiedBy: nasanex01a.na.qualcomm.com (10.52.223.231) To nalasex01b.na.qualcomm.com (10.47.209.197) X-QCInternal: smtphost X-Proofpoint-Virus-Version: vendor=nai engine=6200 definitions=5800 signatures=585085 X-Proofpoint-ORIG-GUID: Jf1M4Vl6VdJJcwp8A4AG2xfEFZWbs8jX X-Proofpoint-GUID: Jf1M4Vl6VdJJcwp8A4AG2xfEFZWbs8jX X-Proofpoint-Virus-Version: vendor=baseguard engine=ICAP:2.0.293,Aquarius:18.0.1039,Hydra:6.0.680,FMLib:17.12.60.29 definitions=2024-09-06_09,2024-09-06_01,2024-09-02_01 X-Proofpoint-Spam-Details: rule=outbound_notspam policy=outbound score=0 priorityscore=1501 mlxscore=0 impostorscore=0 phishscore=0 malwarescore=0 suspectscore=0 bulkscore=0 spamscore=0 mlxlogscore=999 adultscore=0 clxscore=1015 lowpriorityscore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.19.0-2409260000 definitions=main-2410070061 Content-Type: text/plain; charset="utf-8" Add support to capture kernel performance counters for different kernel level operations. These counters collects the information for remote call and copies the information to a buffer shared by user. Collection of DSP performance counters is also added as part of this change. DSP updates the performance information in the metadata which is then copied to a buffer passed by the users. Signed-off-by: Ekansh Gupta --- drivers/misc/fastrpc.c | 143 ++++++++++++++++++++++++++++++++++-- include/uapi/misc/fastrpc.h | 4 +- 2 files changed, 139 insertions(+), 8 deletions(-) diff --git a/drivers/misc/fastrpc.c b/drivers/misc/fastrpc.c index 8e817a763d1d..54a562fc94fb 100644 --- a/drivers/misc/fastrpc.c +++ b/drivers/misc/fastrpc.c @@ -19,6 +19,7 @@ #include #include #include +#include #include #include #include @@ -38,6 +39,7 @@ #define FASTRPC_CTX_MAX (256) #define FASTRPC_INIT_HANDLE 1 #define FASTRPC_DSP_UTILITIES_HANDLE 2 +#define FASTRPC_MAX_STATIC_HANDLE (20) #define FASTRPC_CTXID_MASK (0xFF0) #define INIT_FILELEN_MAX (2 * 1024 * 1024) #define INIT_FILE_NAMELEN_MAX (128) @@ -106,6 +108,9 @@ =20 #define miscdev_to_fdevice(d) container_of(d, struct fastrpc_device, miscd= ev) =20 +#define FASTRPC_KERNEL_PERF_LIST (PERF_KEY_MAX) +#define FASTRPC_DSP_PERF_LIST 12 + static const char *domains[FASTRPC_DEV_MAX] =3D { "adsp", "mdsp", "sdsp", "cdsp", "cdsp1" }; struct fastrpc_phy_page { @@ -229,6 +234,18 @@ struct fastrpc_map { struct kref refcount; }; =20 +struct fastrpc_perf { + u64 count; + u64 flush; + u64 map; + u64 copy; + u64 link; + u64 getargs; + u64 putargs; + u64 invargs; + u64 invoke; +}; + struct fastrpc_invoke_ctx { int nscalars; int nbufs; @@ -237,6 +254,8 @@ struct fastrpc_invoke_ctx { int tgid; u32 sc; u32 *crc; + u64 *perf_kernel; + u64 *perf_dsp; u64 ctxid; u64 msg_sz; struct kref refcount; @@ -251,6 +270,7 @@ struct fastrpc_invoke_ctx { struct fastrpc_invoke_args *args; struct fastrpc_buf_overlap *olaps; struct fastrpc_channel_ctx *cctx; + struct fastrpc_perf *perf; }; =20 struct fastrpc_session_ctx { @@ -308,6 +328,51 @@ struct fastrpc_user { struct mutex mutex; }; =20 +enum fastrpc_perfkeys { + PERF_COUNT =3D 0, + PERF_FLUSH =3D 1, + PERF_MAP =3D 2, + PERF_COPY =3D 3, + PERF_LINK =3D 4, + PERF_GETARGS =3D 5, + PERF_PUTARGS =3D 6, + PERF_INVARGS =3D 7, + PERF_INVOKE =3D 8, + PERF_KEY_MAX =3D 9, +}; + +#define PERF_END ((void)0) + +#define PERF(enb, cnt, ff) \ + {\ + struct timespec64 startT =3D {0};\ + u64 *counter =3D cnt;\ + if (enb && counter) {\ + ktime_get_boottime_ts64(&startT);\ + } \ + ff ;\ + if (enb && counter) {\ + *counter +=3D getnstimediff(&startT);\ + } \ + } + +#define GET_COUNTER(perf_ptr, offset) \ + (perf_ptr !=3D NULL ?\ + (((offset >=3D 0) && (offset < PERF_KEY_MAX)) ?\ + (u64 *)(perf_ptr + offset)\ + : (u64 *)NULL) : (u64 *)NULL) + +static inline s64 getnstimediff(struct timespec64 *start) +{ + s64 ns; + struct timespec64 ts, b; + + ktime_get_boottime_ts64(&ts); + b =3D timespec64_sub(ts, *start); + ns =3D timespec64_to_ns(&b); + return ns; +} + static void fastrpc_free_map(struct kref *ref) { struct fastrpc_map *map; @@ -497,6 +562,7 @@ static void fastrpc_context_free(struct kref *ref) idr_remove(&cctx->ctx_idr, ctx->ctxid >> 4); spin_unlock_irqrestore(&cctx->lock, flags); =20 + kfree(ctx->perf); kfree(ctx->maps); kfree(ctx->olaps); kfree(ctx); @@ -614,6 +680,13 @@ static struct fastrpc_invoke_ctx *fastrpc_context_allo= c( fastrpc_channel_ctx_get(cctx); =20 ctx->crc =3D (u32 *)(uintptr_t)inv2->crc; + ctx->perf_dsp =3D (u64 *)(uintptr_t)inv2->perf_dsp; + ctx->perf_kernel =3D (u64 *)(uintptr_t)inv2->perf_kernel; + if (ctx->perf_kernel) { + ctx->perf =3D kzalloc(sizeof(*(ctx->perf)), GFP_KERNEL); + if (!ctx->perf) + return ERR_PTR(-ENOMEM); + } ctx->sc =3D sc; ctx->retval =3D -1; ctx->pid =3D current->pid; @@ -877,7 +950,8 @@ static int fastrpc_get_meta_size(struct fastrpc_invoke_= ctx *ctx) sizeof(struct fastrpc_invoke_buf) + sizeof(struct fastrpc_phy_page)) * ctx->nscalars + sizeof(u64) * FASTRPC_MAX_FDLIST + - sizeof(u32) * FASTRPC_MAX_CRCLIST; + sizeof(u32) * FASTRPC_MAX_CRCLIST + + sizeof(u32) + sizeof(u64) * FASTRPC_DSP_PERF_LIST; =20 return size; } @@ -951,7 +1025,9 @@ static int fastrpc_get_args(u32 kernel, struct fastrpc= _invoke_ctx *ctx) metalen =3D fastrpc_get_meta_size(ctx); pkt_size =3D fastrpc_get_payload_size(ctx, metalen); =20 + PERF(ctx->perf_kernel, GET_COUNTER((u64 *)ctx->perf, PERF_MAP), err =3D fastrpc_create_maps(ctx); + PERF_END); if (err) return err; =20 @@ -987,6 +1063,7 @@ static int fastrpc_get_args(u32 kernel, struct fastrpc= _invoke_ctx *ctx) continue; =20 if (ctx->maps[i]) { + PERF(ctx->perf_kernel, GET_COUNTER((u64 *)ctx->perf, PERF_MAP), struct vm_area_struct *vma =3D NULL; =20 rpra[i].buf.pv =3D (u64) ctx->args[i].ptr; @@ -1003,9 +1080,9 @@ static int fastrpc_get_args(u32 kernel, struct fastrp= c_invoke_ctx *ctx) pg_end =3D ((ctx->args[i].ptr + len - 1) & PAGE_MASK) >> PAGE_SHIFT; pages[i].size =3D (pg_end - pg_start + 1) * PAGE_SIZE; - + PERF_END); } else { - + PERF(ctx->perf_kernel, GET_COUNTER((u64 *)ctx->perf, PERF_COPY), if (ctx->olaps[oix].offset =3D=3D 0) { rlen -=3D ALIGN(args, FASTRPC_ALIGN) - args; args =3D ALIGN(args, FASTRPC_ALIGN); @@ -1027,12 +1104,14 @@ static int fastrpc_get_args(u32 kernel, struct fast= rpc_invoke_ctx *ctx) pages[i].size =3D (pg_end - pg_start + 1) * PAGE_SIZE; args =3D args + mlen; rlen -=3D mlen; + PERF_END); } =20 if (i < inbufs && !ctx->maps[i]) { void *dst =3D (void *)(uintptr_t)rpra[i].buf.pv; void *src =3D (void *)(uintptr_t)ctx->args[i].ptr; =20 + PERF(ctx->perf_kernel, GET_COUNTER((u64 *)ctx->perf, PERF_COPY), if (!kernel) { if (copy_from_user(dst, (void __user *)src, len)) { @@ -1042,6 +1121,7 @@ static int fastrpc_get_args(u32 kernel, struct fastrp= c_invoke_ctx *ctx) } else { memcpy(dst, src, len); } + PERF_END); } } =20 @@ -1072,9 +1152,9 @@ static int fastrpc_put_args(struct fastrpc_invoke_ctx= *ctx, struct fastrpc_map *mmap =3D NULL; struct fastrpc_invoke_buf *list; struct fastrpc_phy_page *pages; - u64 *fdlist; - u32 *crclist; - int i, inbufs, outbufs, handles; + u64 *fdlist, *perf_dsp_list; + u32 *crclist, *poll; + int i, inbufs, outbufs, handles, perferr; =20 inbufs =3D REMOTE_SCALARS_INBUFS(ctx->sc); outbufs =3D REMOTE_SCALARS_OUTBUFS(ctx->sc); @@ -1083,6 +1163,8 @@ static int fastrpc_put_args(struct fastrpc_invoke_ctx= *ctx, pages =3D fastrpc_phy_page_start(list, ctx->nscalars); fdlist =3D (uint64_t *)(pages + inbufs + outbufs + handles); crclist =3D (u32 *)(fdlist + FASTRPC_MAX_FDLIST); + poll =3D (u32 *)(crclist + FASTRPC_MAX_CRCLIST); + perf_dsp_list =3D (u64 *)(poll + 1); =20 for (i =3D inbufs; i < ctx->nbufs; ++i) { if (!ctx->maps[i]) { @@ -1113,6 +1195,14 @@ static int fastrpc_put_args(struct fastrpc_invoke_ct= x *ctx, return -EFAULT; } =20 + if (ctx->perf_dsp && perf_dsp_list) { + perferr =3D copy_to_user((void __user *)ctx->perf_dsp, + perf_dsp_list, + FASTRPC_DSP_PERF_LIST * sizeof(u64)); + if (perferr) + dev_dbg(fl->sctx->dev, "Warning: failed to copy perf data %d\n", perfer= r); + } + return 0; } =20 @@ -1148,13 +1238,35 @@ static int fastrpc_invoke_send(struct fastrpc_sessi= on_ctx *sctx, =20 } =20 +static void fastrpc_update_invoke_count(u32 handle, u64 *perf_counter, + struct timespec64 *invoket) +{ + /* update invoke count for dynamic handles */ + u64 *invcount, *count; + + if (handle <=3D FASTRPC_MAX_STATIC_HANDLE) + return; + + invcount =3D GET_COUNTER(perf_counter, PERF_INVOKE); + if (invcount) + *invcount +=3D getnstimediff(invoket); + + count =3D GET_COUNTER(perf_counter, PERF_COUNT); + if (count) + *count++; +} + static int fastrpc_internal_invoke(struct fastrpc_user *fl, u32 kernel, s= truct fastrpc_invoke_v2 *inv2) { struct fastrpc_invoke_ctx *ctx =3D NULL; struct fastrpc_buf *buf, *b; struct fastrpc_invoke inv; u32 handle, sc; - int err =3D 0; + int err =3D 0, perferr =3D 0; + struct timespec64 invoket =3D {0}; + + if (inv2->perf_kernel) + ktime_get_boottime_ts64(&invoket); =20 if (!fl->sctx) return -EINVAL; @@ -1174,14 +1286,18 @@ static int fastrpc_internal_invoke(struct fastrpc_u= ser *fl, u32 kernel, struct if (IS_ERR(ctx)) return PTR_ERR(ctx); =20 + PERF(ctx->perf_kernel, GET_COUNTER((u64 *)ctx->perf, PERF_GETARGS), err =3D fastrpc_get_args(kernel, ctx); + PERF_END); if (err) goto bail; =20 /* make sure that all CPU memory writes are seen by DSP */ dma_wmb(); + PERF(ctx->perf_kernel, GET_COUNTER((u64 *)ctx->perf, PERF_LINK), /* Send invoke buffer to remote dsp */ err =3D fastrpc_invoke_send(fl->sctx, ctx, kernel, handle); + PERF_END); if (err) goto bail; =20 @@ -1197,8 +1313,10 @@ static int fastrpc_internal_invoke(struct fastrpc_us= er *fl, u32 kernel, struct =20 /* make sure that all memory writes by DSP are seen by CPU */ dma_rmb(); + PERF(ctx->perf_kernel, GET_COUNTER((u64 *)ctx->perf, PERF_PUTARGS), /* populate all the output buffers with results */ err =3D fastrpc_put_args(ctx, kernel); + PERF_END); if (err) goto bail; =20 @@ -1209,6 +1327,17 @@ static int fastrpc_internal_invoke(struct fastrpc_us= er *fl, u32 kernel, struct =20 bail: if (err !=3D -ERESTARTSYS && err !=3D -ETIMEDOUT) { + if (ctx) { + if (ctx->perf_kernel && !err) + fastrpc_update_invoke_count(handle, (u64 *)ctx->perf, &invoket); + if (ctx->perf_kernel && ctx->perf && ctx->perf_kernel) { + perferr =3D copy_to_user((void __user *)ctx->perf_kernel, + ctx->perf, + FASTRPC_KERNEL_PERF_LIST * sizeof(u64)); + if (perferr) + dev_dbg(fl->sctx->dev, "Warning: failed to copy perf data %d\n", perf= err); + } + } /* We are done with this compute context */ spin_lock(&fl->lock); list_del(&ctx->node); diff --git a/include/uapi/misc/fastrpc.h b/include/uapi/misc/fastrpc.h index 406b80555d41..1edc7c04b171 100644 --- a/include/uapi/misc/fastrpc.h +++ b/include/uapi/misc/fastrpc.h @@ -84,7 +84,9 @@ struct fastrpc_invoke { struct fastrpc_invoke_v2 { struct fastrpc_invoke inv; __u64 crc; - __u32 reserved[16]; + __u64 perf_kernel; + __u64 perf_dsp; + __u32 reserved[12]; }; =20 struct fastrpc_init_create { --=20 2.34.1