From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from dpdk.org (dpdk.org [92.243.14.124]) by inbox.dpdk.org (Postfix) with ESMTP id C49B9A0350; Fri, 26 Jun 2020 13:03:24 +0200 (CEST) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id B11F21BEA9; Fri, 26 Jun 2020 13:03:23 +0200 (CEST) Received: from us-smtp-delivery-1.mimecast.com (us-smtp-1.mimecast.com [207.211.31.81]) by dpdk.org (Postfix) with ESMTP id 30AA81BE98 for ; Fri, 26 Jun 2020 13:03:22 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1593169401; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:autocrypt:autocrypt; bh=Z/g6ln4Sei0l8rGD1v+pFcQXUgCrN99r21FmBVtRiFg=; b=etzxqpKsRw6Y9osOpxbxhbagCGHnf0aMDrYsCpsC+/Z2xNpDi4tbq2BtJnQFP11nd7mMx2 nANOfvmJ4PikmeorDbX4QC6KM6G+K+IjzUDNC9LoVv5er5aZeSNSF38zS9luBNr44PnGoC jvkfTv6CoGicfuGThU0M3/dz3u2doz0= Received: from mail-wr1-f72.google.com (mail-wr1-f72.google.com [209.85.221.72]) (Using TLS) by relay.mimecast.com with ESMTP id us-mta-351-zZfm6zjcMmWvdAqKTG9ipg-1; Fri, 26 Jun 2020 07:03:19 -0400 X-MC-Unique: zZfm6zjcMmWvdAqKTG9ipg-1 Received: by mail-wr1-f72.google.com with SMTP id o12so10191660wrj.23 for ; Fri, 26 Jun 2020 04:03:19 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:subject:to:references:from:autocrypt:message-id :date:user-agent:mime-version:in-reply-to:content-language :content-transfer-encoding; bh=Z/g6ln4Sei0l8rGD1v+pFcQXUgCrN99r21FmBVtRiFg=; b=nMiAQDEmwXP55XApEgt6/xU46yEwO/42IyCKrdtmJ3Aoff+3urkOxkEDClHa+/i+Bw 9Bx0bZ5LGO74GIGfIjwyHVRDEPEM1MgHkUTyeZlVIMVoHJyameEh5l4EToQBWaO/SBG8 ltd3fL5JCSXg7xWDqjGYQDOcOwTrwwXRqvastHyaui04y2zOC3vUljyJVkeduRsppvN1 01znibMfPFGe0kNiPF40++YNB1uy6BOgSr4xH1uPOKyVqvnUrS4leQXr0QU1btUWi1Ss AGgQ4vqvTBX972hgT05IOdjkDhc/XRmnZI/nzSckLphJhO9ewzmajWnU4wxG35okhbTC IYng== X-Gm-Message-State: AOAM533VTlI/f9FU4GnIc0g4FFR5kdbz3Bu/QXa0fwWc1nuTBNItKYow 4PLSwYG/zGC6eS3R2b7e04ynuPRjWqvWAs3TOED1gYIuAcWmFawpdmntyZlWIbIbXNAV6JZ4ZBW Eetk= X-Received: by 2002:adf:e7c8:: with SMTP id e8mr3130284wrn.20.1593169396676; Fri, 26 Jun 2020 04:03:16 -0700 (PDT) X-Google-Smtp-Source: ABdhPJxEe9BMs2FJqokrcx+9Q4i5rvqlxJgV7+qqULZvW1kq5rjet9xQAdHQJ15FCYV7pvCLzvXarg== X-Received: by 2002:adf:e7c8:: with SMTP id e8mr3130247wrn.20.1593169396350; Fri, 26 Jun 2020 04:03:16 -0700 (PDT) Received: from amorenoz.users.ipa.redhat.com ([92.176.159.3]) by smtp.gmail.com with ESMTPSA id f1sm16574472wmj.12.2020.06.26.04.03.15 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Fri, 26 Jun 2020 04:03:15 -0700 (PDT) To: Maxime Coquelin , matan@mellanox.com, xiao.w.wang@intel.com, zhihong.wang@intel.com, chenbo.xia@intel.com, david.marchand@redhat.com, viacheslavo@mellanox.com, hemant.agrawal@nxp.com, sachin.saxena@nxp.com, grive@u256.net, dev@dpdk.org References: <20200624122701.1369327-1-maxime.coquelin@redhat.com> <20200624122701.1369327-10-maxime.coquelin@redhat.com> From: Adrian Moreno Autocrypt: addr=amorenoz@redhat.com; prefer-encrypt=mutual; keydata= mQENBF1syNUBCADQ9dk3fDMxOZ/+OQpmbanpodYxEv8IRtDz8PXw8YX7UyGfozOpLjQ8Fftj ZxuubYNbt2QVbSgviFilFdNWu2eTnN/JaGtfhmTOLPVoakkPHZF8lbgImMoch7L0fH8wN2IM KPxQyPNlX+K9FD5brHsV1lfe1TwAxmhcvLW8yNrVq+9eDIDykxc7tH4exIqXgZroahGxMHKy c8Ti2kJka/t6pDfRaY0J+6J7I1nrn6GXXSMNA45EH8+0N/QlcXhP3rfftnoPeVmpjswzvJqY FNjf/Q5VPLx7RX0Qx+y8mMB2JcChV5Bl7D7x5EUbItj6+Sy7QfOgCtPegk9HSrBCNYaLABEB AAG0I0FkcmlhbiBNb3Jlbm8gPGFtb3Jlbm96QHJlZGhhdC5jb20+iQFUBBMBCAA+FiEEogUD gihhmbOPHy26d5C5fbYeFsUFAl1syNUCGwMFCQHhM4AFCwkIBwIGFQoJCAsCBBYCAwECHgEC F4AACgkQd5C5fbYeFsX7qwgArGHSkX+ILNcujkVzjTG4OtkpJMPFlkn/1PxSEKD0jLuzx14B COzpg/Mqj3Re/QBuOas+ci9bsUA0/2nORtmmEBvzDOJpR5FH1jaGCx8USlY4WM6QqEDNZgTw hsy9KhjFzFjMk+oo3HyItXA+Uq9yrRBTjNBGTXxezMRcMuUZ4MIAfY0IRBglL2BufiuL43jD BvTENNFLoQ/wFV7qkFWSkv+8IjTsxr7M6XUo1QLd29Hn0dvwssN579HL1+BP46i2REpzeBEG L75iVChi+YnIQQNMJ9NYarVabZx4Y1Gn8+7B/1SNArDV+IDgnYgt7E58otoV2Ap310dmtuvE VbxGpbkBDQRdbMjVAQgAqyp9oA7WDu7/Y9T4Ommt69iZx8os7shUIfdgPEy5xrcPn6qGwN1/ HQ4j8nWfBG9uuX1X0RXUZIUEtYTxtED4yaCQMTqDUf9cBAwAA2mYxBfoiNYx8YqxM+sT0/J4 2qmDd+y+20UR4yzHE8AmIbspTzDFIJDAi+jKSR8F355z0sfW7CIMDC4ZWrPsskjEy1YN/U10 r6tRRH1kNyrCSbTG0d9MtcQO58h7DLXuzUhErB+BtG52A04t5cweIJTJC+koV5XPeilzlHnm RFoj0ncruGa9Odns21BNt3cy9wLfK+aUnWuAB1uc6bJGQPiAwjkilz7g7MBRUuIQ2Zt7HGLc SwARAQABiQE8BBgBCAAmFiEEogUDgihhmbOPHy26d5C5fbYeFsUFAl1syNUCGwwFCQHhM4AA CgkQd5C5fbYeFsUlSwf8CH+u/IXaE7WeWxwFkMaORfW8cM4q0xrL3M6yRGuQNW+kMjnrvK9U J9G+L1/5uTRbDQ/4LdoKqize8LjehA+iF6ba4t9Npikh8fLKWgaJfQ/hPhH4C3O5gWPOLTW6 ylGxiuER4CdFwQIoAMMslhFA7G+teeOKBq36E+1+zrybI6Xy1UBSlpDK9j4CtTnMQejjuSQb Qhle+l8VroaUHq869wjAhRHHhqmtJKggI+OvzgQpDIwfHIDypb1BuKydi2W6cVYEALUYyCLS dTBDhzj8zR5tPCsga8J7+TclQzkWOiI2C6ZtiWrMsL/Uym3uXk5nsoc7lSj7yLd/MrBRhYfP JQ== Message-ID: <5e84d1dd-52d1-25f6-ca99-e73411fc247c@redhat.com> Date: Fri, 26 Jun 2020 13:03:14 +0200 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.8.0 MIME-Version: 1.0 In-Reply-To: <20200624122701.1369327-10-maxime.coquelin@redhat.com> Content-Language: en-US X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Subject: Re: [dpdk-dev] [PATCH v2 09/14] vhost: use linked-list for vDPA devices X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" On 6/24/20 2:26 PM, Maxime Coquelin wrote: > There is no more notion of device ID outside of vdpa.c. > We can now move from array to linked-list model for keeping > track of the vDPA devices. > > There is no point in using array here, as all vDPA API are > used from the control path, so no performance concerns. > > Signed-off-by: Maxime Coquelin > --- > lib/librte_vhost/rte_vdpa.h | 1 + > lib/librte_vhost/vdpa.c | 134 +++++++++++++++++++----------------- > 2 files changed, 70 insertions(+), 65 deletions(-) > > diff --git a/lib/librte_vhost/rte_vdpa.h b/lib/librte_vhost/rte_vdpa.h > index da8ee33ab1..dbdc273702 100644 > --- a/lib/librte_vhost/rte_vdpa.h > +++ b/lib/librte_vhost/rte_vdpa.h > @@ -105,6 +105,7 @@ struct rte_vdpa_dev_ops { > * vdpa device structure includes device address and device operations. > */ > struct rte_vdpa_device { > + TAILQ_ENTRY(rte_vdpa_device) next; > /** Generic device information */ > struct rte_device *device; > /** vdpa device operations */ > diff --git a/lib/librte_vhost/vdpa.c b/lib/librte_vhost/vdpa.c > index d7965b8854..dc2a138009 100644 > --- a/lib/librte_vhost/vdpa.c > +++ b/lib/librte_vhost/vdpa.c > @@ -9,35 +9,54 @@ > */ > > #include > +#include > > #include > #include > +#include > +#include > + I'm curious why we need both sys/queue.h and rte_tailq.h > #include "rte_vdpa.h" > #include "vhost.h" > > -static struct rte_vdpa_device vdpa_devices[MAX_VHOST_DEVICE]; > +/** Double linked list of vDPA devices. */ > +TAILQ_HEAD(vdpa_device_list, rte_vdpa_device); > + > +static struct vdpa_device_list vdpa_device_list = > + TAILQ_HEAD_INITIALIZER(vdpa_device_list); > +static rte_spinlock_t vdpa_device_list_lock = RTE_SPINLOCK_INITIALIZER; > static uint32_t vdpa_device_num; > > > -struct rte_vdpa_device * > -rte_vdpa_find_device_by_name(const char *name) > +/* Unsafe, needs to be called with vdpa_device_list_lock held */ > +static struct rte_vdpa_device * > +__vdpa_find_device_by_name(const char *name) > { > - struct rte_vdpa_device *dev; > - int i; > + struct rte_vdpa_device *dev, *ret = NULL; > > if (name == NULL) > return NULL; > > - for (i = 0; i < MAX_VHOST_DEVICE; ++i) { > - dev = &vdpa_devices[i]; > - if (dev->ops == NULL) > - continue; > - > - if (strcmp(dev->device->name, name) == 0) > - return dev; > + TAILQ_FOREACH(dev, &vdpa_device_list, next) { > + if (strcmp(dev->device->name, name) == 0) { > + ret = dev; > + break; > + } > } > > - return NULL; > + return ret; > +} > + > +struct rte_vdpa_device * > +rte_vdpa_find_device_by_name(const char *name) > +{ > + struct rte_vdpa_device *dev; > + > + rte_spinlock_lock(&vdpa_device_list_lock); > + dev = __vdpa_find_device_by_name(name); > + rte_spinlock_unlock(&vdpa_device_list_lock); > + > + return dev; > } > > struct rte_device * > @@ -54,52 +73,52 @@ rte_vdpa_register_device(struct rte_device *rte_dev, > struct rte_vdpa_dev_ops *ops) > { > struct rte_vdpa_device *dev; > - int i; > > - if (vdpa_device_num >= MAX_VHOST_DEVICE || ops == NULL) > + if (ops == NULL) > return NULL; > > - for (i = 0; i < MAX_VHOST_DEVICE; i++) { > - dev = &vdpa_devices[i]; > - if (dev->ops == NULL) > - continue; > - > - if (dev->device == rte_dev) > - return NULL; > - } > - > - for (i = 0; i < MAX_VHOST_DEVICE; i++) { > - if (vdpa_devices[i].ops == NULL) > - break; > + rte_spinlock_lock(&vdpa_device_list_lock); > + /* Check the device hasn't been register already */ > + dev = __vdpa_find_device_by_name(rte_dev->name); > + if (dev) { > + dev = NULL; > + goto out_unlock; > } > > - if (i == MAX_VHOST_DEVICE) > - return NULL; > + dev = rte_zmalloc(NULL, sizeof(*dev), 0); > + if (!dev) > + goto out_unlock; > > - dev = &vdpa_devices[i]; > dev->device = rte_dev; > dev->ops = ops; > + TAILQ_INSERT_TAIL(&vdpa_device_list, dev, next); > vdpa_device_num++; > +out_unlock: > + rte_spinlock_unlock(&vdpa_device_list_lock); > > return dev; > } > > int > -rte_vdpa_unregister_device(struct rte_vdpa_device *vdev) > +rte_vdpa_unregister_device(struct rte_vdpa_device *dev) > { > - int i; > + struct rte_vdpa_device *cur_dev, *tmp_dev; > + int ret = -1; > > - for (i = 0; i < MAX_VHOST_DEVICE; i++) { > - if (vdev != &vdpa_devices[i]) > + rte_spinlock_lock(&vdpa_device_list_lock); > + TAILQ_FOREACH_SAFE(cur_dev, &vdpa_device_list, next, tmp_dev) { > + if (dev != cur_dev) > continue; > > - memset(vdev, 0, sizeof(struct rte_vdpa_device)); > + TAILQ_REMOVE(&vdpa_device_list, dev, next); > + rte_free(dev); > vdpa_device_num--; > - > - return 0; > + ret = 0; > + break; > } > + rte_spinlock_lock(&vdpa_device_list_lock); > You mean rte_spinlock_unlock? > - return -1; > + return ret; > } > > int > @@ -246,19 +265,6 @@ rte_vdpa_reset_stats(struct rte_vdpa_device *dev, uint16_t qid) > return dev->ops->reset_stats(dev, qid); > } > > -static uint16_t > -vdpa_dev_to_id(const struct rte_vdpa_device *dev) > -{ > - if (dev == NULL) > - return MAX_VHOST_DEVICE; > - > - if (dev < &vdpa_devices[0] || > - dev >= &vdpa_devices[MAX_VHOST_DEVICE]) > - return MAX_VHOST_DEVICE; > - > - return (uint16_t)(dev - vdpa_devices); > -} > - > static int > vdpa_dev_match(struct rte_vdpa_device *dev, > const struct rte_device *rte_dev) > @@ -278,24 +284,22 @@ vdpa_find_device(const struct rte_vdpa_device *start, rte_vdpa_cmp_t cmp, > struct rte_device *rte_dev) > { > struct rte_vdpa_device *dev; > - uint16_t idx; > > - if (start != NULL) > - idx = vdpa_dev_to_id(start) + 1; > + rte_spinlock_lock(&vdpa_device_list_lock); > + if (start == NULL) > + dev = TAILQ_FIRST(&vdpa_device_list); > else > - idx = 0; > - for (; idx < MAX_VHOST_DEVICE; idx++) { > - dev = &vdpa_devices[idx]; > - /* > - * ToDo: Certainly better to introduce a state field, > - * but rely on ops being set for now. > - */ > - if (dev->ops == NULL) > - continue; > + dev = TAILQ_NEXT(start, next); > + > + while (dev != NULL) { > if (cmp(dev, rte_dev) == 0) > - return dev; > + break; > + > + dev = TAILQ_NEXT(dev, next); > } > - return NULL; > + rte_spinlock_unlock(&vdpa_device_list_lock); > + > + return dev; > } > > static void * > -- Adrián Moreno