From nobody Thu Nov 28 02:34:39 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 EC5F718C90A; Mon, 7 Oct 2024 08:45:48 +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=1728290750; cv=none; b=TNjIbkVs9PqGbfGQDBAp+MvAjtkzhOFswJMPWjTrYYKny3OPXRP7utF8h++39g5q8d+pyqtDfGVBVj0h+PooluXz5oX9tKKefsUyEseUe0QBFEP1QpVaJYYbvdzWZ3P5oV1tGFposFIQ6wqyBVByWexSEJnlT7gFzp/uZgjlnmc= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1728290750; c=relaxed/simple; bh=UG72Cpzkki87rQALIJWboNyc6S7PCXSW4dMpJctdtPw=; h=From:To:CC:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version:Content-Type; b=FkAVgkRJbQo7btae8jMNb9MkCasDDj8kcc5MityonmpezbjJP2V1NVy09uZF0tHidmFSBxjziE+LJ1cpp1kbfCa4iwq/9Ah7bRlbwhxj5onEP5RaYQ9hf5wZzDH6pTFGOqT0QcIxxAO0uq8VmT0uTHTeIioJiv+p5czKJ9D5mzg= 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=Arln9c0R; 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="Arln9c0R" Received: from pps.filterd (m0279864.ppops.net [127.0.0.1]) by mx0a-0031df01.pphosted.com (8.18.1.2/8.18.1.2) with ESMTP id 4974TfF0021858; Mon, 7 Oct 2024 08:45:42 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= iiGu/T4YMqs9vu7zT6kk9SjYjwUUm6sCok1Z5LDxw9s=; b=Arln9c0RSPiHsmI5 SDAhItocFrrTWqOM1++5O50zknnBN0sWleYW4HCYLmkk26997xxmwDOc8T9WmkQL +XoB1XcAKCX4flOH3LyabaDf4E2O8SC+lQqwsXNvrYEMs22OmsAhpQ26cacZSEDG 4XnZwBVLi/Owszui56G/TxCmBEY5/pyrYoyREYxZXSEhuwnUFfjbGbOEXER2AbKx Fk/WLKwnxUHpKSjIgQjBUOiRRIPPUbHzEzMcRE4ihOeWa4mNcJF7MIyzTHQ0tkGz RmZJpbyETJQJOBsgQvMYxw72mhXsJ+f1bmoeaBtuZBnSBeLjHqNxj0WJiphnBfdH nqHCaA== Received: from nalasppmta05.qualcomm.com (Global_NAT1.qualcomm.com [129.46.96.20]) by mx0a-0031df01.pphosted.com (PPS) with ESMTPS id 422xr5kgjj-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 07 Oct 2024 08:45:42 +0000 (GMT) Received: from nalasex01b.na.qualcomm.com (nalasex01b.na.qualcomm.com [10.47.209.197]) by NALASPPMTA05.qualcomm.com (8.18.1.2/8.18.1.2) with ESMTPS id 4978jfCQ010040 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 7 Oct 2024 08:45:41 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 01:45:39 -0700 From: Ekansh Gupta To: , CC: , , , , , Subject: [PATCH v1 4/4] misc: fastrpc: Add polling mode support for fastRPC driver Date: Mon, 7 Oct 2024 14:15:18 +0530 Message-ID: <20241007084518.3649876-5-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-GUID: 6JLd1kRHvyREZ8gnek10FIlMq_Qsjmsf X-Proofpoint-ORIG-GUID: 6JLd1kRHvyREZ8gnek10FIlMq_Qsjmsf 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 spamscore=0 suspectscore=0 clxscore=1015 mlxlogscore=999 priorityscore=1501 impostorscore=0 adultscore=0 mlxscore=0 phishscore=0 malwarescore=0 lowpriorityscore=0 bulkscore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.19.0-2409260000 definitions=main-2410070061 Content-Type: text/plain; charset="utf-8" For any remote call to DSP, after sending an invocation message, fastRPC driver waits for glink response and during this time the CPU can go into low power modes. Adding a polling mode support with which fastRPC driver will poll continuously on a memory after sending a message to remote subsystem which will eliminate CPU wakeup and scheduling latencies and reduce fastRPC overhead. With this change, DSP always sends a glink response which will get ignored if polling mode didn't time out. Signed-off-by: Ekansh Gupta --- drivers/misc/fastrpc.c | 127 +++++++++++++++++++++++++++++++++--- include/uapi/misc/fastrpc.h | 3 +- 2 files changed, 121 insertions(+), 9 deletions(-) diff --git a/drivers/misc/fastrpc.c b/drivers/misc/fastrpc.c index c008fcd95e15..a25673c22db8 100644 --- a/drivers/misc/fastrpc.c +++ b/drivers/misc/fastrpc.c @@ -111,6 +111,19 @@ #define FASTRPC_KERNEL_PERF_LIST (PERF_KEY_MAX) #define FASTRPC_DSP_PERF_LIST 12 =20 +/* Poll response number from remote processor for call completion */ +#define FASTRPC_POLL_RESPONSE (0xdecaf) +/* timeout in us for polling until memory barrier */ +#define FASTRPC_POLL_TIME_MEM_UPDATE (500) + +/* Response types supported for RPC calls */ +enum fastrpc_response_flags { + /* normal job completion glink response */ + NORMAL_RESPONSE =3D 0, + /* process updates poll memory instead of glink response */ + POLL_MODE =3D 1, +}; + static const char *domains[FASTRPC_DEV_MAX] =3D { "adsp", "mdsp", "sdsp", "cdsp", "cdsp1" }; struct fastrpc_phy_page { @@ -258,6 +271,12 @@ struct fastrpc_invoke_ctx { u64 *perf_dsp; u64 ctxid; u64 msg_sz; + /* Threads poll for specified timeout and fall back to glink wait */ + u64 poll_timeout; + /* work done status flag */ + bool is_work_done; + /* response flags from remote processor */ + enum fastrpc_response_flags rsp_flags; struct kref refcount; struct list_head node; /* list of ctxs */ struct completion work; @@ -682,6 +701,7 @@ static struct fastrpc_invoke_ctx *fastrpc_context_alloc( 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; + ctx->poll_timeout =3D (u64)inv2->poll_timeout; if (ctx->perf_kernel) { ctx->perf =3D kzalloc(sizeof(*(ctx->perf)), GFP_KERNEL); if (!ctx->perf) @@ -692,6 +712,8 @@ static struct fastrpc_invoke_ctx *fastrpc_context_alloc( ctx->pid =3D current->pid; ctx->tgid =3D user->tgid; ctx->cctx =3D cctx; + ctx->rsp_flags =3D NORMAL_RESPONSE; + ctx->is_work_done =3D false; init_completion(&ctx->work); INIT_WORK(&ctx->put_work, fastrpc_context_put_wq); =20 @@ -1256,6 +1278,87 @@ static void fastrpc_update_invoke_count(u32 handle, = u64 *perf_counter, *count++; } =20 +static int poll_for_remote_response(struct fastrpc_invoke_ctx *ctx, u64 ti= meout) +{ + int err =3D -EIO, i, j; + u32 sc =3D ctx->sc; + struct fastrpc_invoke_buf *list; + struct fastrpc_phy_page *pages; + u64 *fdlist =3D NULL; + u32 *crclist =3D NULL, *poll =3D NULL; + unsigned int inbufs, outbufs, handles; + + /* calculate poll memory location */ + inbufs =3D REMOTE_SCALARS_INBUFS(sc); + outbufs =3D REMOTE_SCALARS_OUTBUFS(sc); + handles =3D REMOTE_SCALARS_INHANDLES(sc) + REMOTE_SCALARS_OUTHANDLES(sc); + list =3D fastrpc_invoke_buf_start(ctx->rpra, ctx->nscalars); + pages =3D fastrpc_phy_page_start(list, ctx->nscalars); + fdlist =3D (u64 *)(pages + inbufs + outbufs + handles); + crclist =3D (u32 *)(fdlist + FASTRPC_MAX_FDLIST); + poll =3D (u32 *)(crclist + FASTRPC_MAX_CRCLIST); + + /* poll on memory for DSP response. Return failure on timeout */ + for (i =3D 0, j =3D 0; i < timeout; i++, j++) { + if (*poll =3D=3D FASTRPC_POLL_RESPONSE) { + err =3D 0; + ctx->is_work_done =3D true; + ctx->retval =3D 0; + break; + } + if (j =3D=3D FASTRPC_POLL_TIME_MEM_UPDATE) { + /* make sure that all poll memory writes by DSP are seen by CPU */ + dma_rmb(); + j =3D 0; + } + udelay(1); + } + return err; +} + +static inline int fastrpc_wait_for_response(struct fastrpc_invoke_ctx *ctx, + u32 kernel) +{ + int err =3D 0; + + if (kernel) { + if (!wait_for_completion_timeout(&ctx->work, 10 * HZ)) + err =3D -ETIMEDOUT; + } else { + err =3D wait_for_completion_interruptible(&ctx->work); + } + + return err; +} + +static int fastrpc_wait_for_completion(struct fastrpc_invoke_ctx *ctx, + u32 kernel) +{ + int err; + + do { + switch (ctx->rsp_flags) { + case NORMAL_RESPONSE: + err =3D fastrpc_wait_for_response(ctx, kernel); + if (err || ctx->is_work_done) + return err; + break; + case POLL_MODE: + err =3D poll_for_remote_response(ctx, ctx->poll_timeout); + /* If polling timed out, move to normal response mode */ + if (err) + ctx->rsp_flags =3D NORMAL_RESPONSE; + break; + default: + err =3D -EBADR; + dev_dbg(ctx->fl->sctx->dev, "unsupported response type:0x%x\n", ctx->rs= p_flags); + break; + } + } while (!ctx->is_work_done); + + return err; +} + static int fastrpc_internal_invoke(struct fastrpc_user *fl, u32 kernel, s= truct fastrpc_invoke_v2 *inv2) { struct fastrpc_invoke_ctx *ctx =3D NULL; @@ -1301,16 +1404,22 @@ static int fastrpc_internal_invoke(struct fastrpc_u= ser *fl, u32 kernel, struct if (err) goto bail; =20 - if (kernel) { - if (!wait_for_completion_timeout(&ctx->work, 10 * HZ)) - err =3D -ETIMEDOUT; - } else { - err =3D wait_for_completion_interruptible(&ctx->work); - } + if (ctx->poll_timeout !=3D 0 && handle > FASTRPC_MAX_STATIC_HANDLE && + fl->cctx->domain_id =3D=3D CDSP_DOMAIN_ID && + fl->pd =3D=3D USER_PD) + ctx->rsp_flags =3D POLL_MODE; =20 + err =3D fastrpc_wait_for_completion(ctx, kernel); if (err) goto bail; =20 + if (!ctx->is_work_done) { + err =3D -ETIMEDOUT; + dev_dbg(fl->sctx->dev, "Invalid workdone state for handle 0x%x, sc 0x%x\= n", + handle, sc); + goto bail; + } + /* 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), @@ -2627,12 +2736,14 @@ static int fastrpc_rpmsg_callback(struct rpmsg_devi= ce *rpdev, void *data, ctx =3D idr_find(&cctx->ctx_idr, ctxid); spin_unlock_irqrestore(&cctx->lock, flags); =20 + /* Ignore this failure as context returned will be NULL for polling mode = */ if (!ctx) { - dev_err(&rpdev->dev, "No context ID matches response\n"); - return -ENOENT; + dev_dbg(&rpdev->dev, "No context ID matches response\n"); + return 0; } =20 ctx->retval =3D rsp->retval; + ctx->is_work_done =3D true; complete(&ctx->work); =20 /* diff --git a/include/uapi/misc/fastrpc.h b/include/uapi/misc/fastrpc.h index 1edc7c04b171..dde684554011 100644 --- a/include/uapi/misc/fastrpc.h +++ b/include/uapi/misc/fastrpc.h @@ -86,7 +86,8 @@ struct fastrpc_invoke_v2 { __u64 crc; __u64 perf_kernel; __u64 perf_dsp; - __u32 reserved[12]; + __u64 poll_timeout; + __u32 reserved[10]; }; =20 struct fastrpc_init_create { --=20 2.34.1