From patchwork Mon Dec 16 19:01:16 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Logan Gunthorpe X-Patchwork-Id: 11295055 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 3EA971593 for ; Mon, 16 Dec 2019 19:01:37 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 2625D2064B for ; Mon, 16 Dec 2019 19:01:37 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726454AbfLPTB0 (ORCPT ); Mon, 16 Dec 2019 14:01:26 -0500 Received: from ale.deltatee.com ([207.54.116.67]:34932 "EHLO ale.deltatee.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1725836AbfLPTBZ (ORCPT ); Mon, 16 Dec 2019 14:01:25 -0500 Received: from cgy1-donard.priv.deltatee.com ([172.16.1.31]) by ale.deltatee.com with esmtps (TLS1.3:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1igvcR-0005jD-3e; Mon, 16 Dec 2019 12:01:23 -0700 Received: from gunthorp by cgy1-donard.priv.deltatee.com with local (Exim 4.92) (envelope-from ) id 1igvcQ-0005Za-86; Mon, 16 Dec 2019 12:01:22 -0700 From: Logan Gunthorpe To: linux-kernel@vger.kernel.org, dmaengine@vger.kernel.org, Vinod Koul Cc: Dan Williams , Dave Jiang , Kit Chow , Logan Gunthorpe Date: Mon, 16 Dec 2019 12:01:16 -0700 Message-Id: <20191216190120.21374-2-logang@deltatee.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20191216190120.21374-1-logang@deltatee.com> References: <20191216190120.21374-1-logang@deltatee.com> MIME-Version: 1.0 X-SA-Exim-Connect-IP: 172.16.1.31 X-SA-Exim-Rcpt-To: linux-kernel@vger.kernel.org, dmaengine@vger.kernel.org, vkoul@kernel.org, dan.j.williams@intel.com, dave.jiang@intel.com, kchow@gigaio.com, logang@deltatee.com X-SA-Exim-Mail-From: gunthorp@deltatee.com X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on ale.deltatee.com X-Spam-Level: X-Spam-Status: No, score=-6.7 required=5.0 tests=ALL_TRUSTED,BAYES_00, MYRULES_NO_TEXT autolearn=no autolearn_force=no version=3.4.2 Subject: [PATCH 1/5] dmaengine: Store module owner in dma_device struct X-SA-Exim-Version: 4.2.1 (built Wed, 08 May 2019 21:11:16 +0000) X-SA-Exim-Scanned: Yes (on ale.deltatee.com) Sender: dmaengine-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: dmaengine@vger.kernel.org dma_chan_to_owner() dereferences the driver from the struct device to obtain the owner and call module_[get|put](). However, if the backing device is unbound before the dma_device is unregistered, the driver will be cleared and this will cause a NULL pointer dereference. Instead, store a pointer to the owner module in the dma_device struct so the module reference can be properly put when the channel is put, even if the backing device was destroyed first. This change helps to support a safer unbind of DMA engines. If the dma_device is unregistered in the driver's remove function, there's no guarantee that there are no existing clients and a users action may trigger the WARN_ONCE in dma_async_device_unregister() which is unlikely to leave the system in a consistent state. Instead, a better approach is to allow the backing driver to go away and fail any subsequent requests to it. Signed-off-by: Logan Gunthorpe --- drivers/dma/dmaengine.c | 4 +++- include/linux/dmaengine.h | 2 ++ 2 files changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c index 03ac4b96117c..4b604086b1b3 100644 --- a/drivers/dma/dmaengine.c +++ b/drivers/dma/dmaengine.c @@ -179,7 +179,7 @@ __dma_device_satisfies_mask(struct dma_device *device, static struct module *dma_chan_to_owner(struct dma_chan *chan) { - return chan->device->dev->driver->owner; + return chan->device->owner; } /** @@ -919,6 +919,8 @@ int dma_async_device_register(struct dma_device *device) return -EIO; } + device->owner = device->dev->driver->owner; + if (dma_has_cap(DMA_MEMCPY, device->cap_mask) && !device->device_prep_dma_memcpy) { dev_err(device->dev, "Device claims capability %s, but op is not defined\n", diff --git a/include/linux/dmaengine.h b/include/linux/dmaengine.h index 8fcdee1c0cf9..13aa0abb71de 100644 --- a/include/linux/dmaengine.h +++ b/include/linux/dmaengine.h @@ -674,6 +674,7 @@ struct dma_filter { * @fill_align: alignment shift for memset operations * @dev_id: unique device ID * @dev: struct device reference for dma mapping api + * @owner: owner module (automatically set based on the provided dev) * @src_addr_widths: bit mask of src addr widths the device supports * Width is specified in bytes, e.g. for a device supporting * a width of 4 the mask should have BIT(4) set. @@ -737,6 +738,7 @@ struct dma_device { int dev_id; struct device *dev; + struct module *owner; u32 src_addr_widths; u32 dst_addr_widths; From patchwork Mon Dec 16 19:01:17 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Logan Gunthorpe X-Patchwork-Id: 11295061 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id BF06514B7 for ; Mon, 16 Dec 2019 19:01:47 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 97BB3206E0 for ; Mon, 16 Dec 2019 19:01:47 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726695AbfLPTBZ (ORCPT ); Mon, 16 Dec 2019 14:01:25 -0500 Received: from ale.deltatee.com ([207.54.116.67]:34938 "EHLO ale.deltatee.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726427AbfLPTBZ (ORCPT ); Mon, 16 Dec 2019 14:01:25 -0500 Received: from cgy1-donard.priv.deltatee.com ([172.16.1.31]) by ale.deltatee.com with esmtps (TLS1.3:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1igvcR-0005jE-3f; Mon, 16 Dec 2019 12:01:24 -0700 Received: from gunthorp by cgy1-donard.priv.deltatee.com with local (Exim 4.92) (envelope-from ) id 1igvcQ-0005Zd-C9; Mon, 16 Dec 2019 12:01:22 -0700 From: Logan Gunthorpe To: linux-kernel@vger.kernel.org, dmaengine@vger.kernel.org, Vinod Koul Cc: Dan Williams , Dave Jiang , Kit Chow , Logan Gunthorpe Date: Mon, 16 Dec 2019 12:01:17 -0700 Message-Id: <20191216190120.21374-3-logang@deltatee.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20191216190120.21374-1-logang@deltatee.com> References: <20191216190120.21374-1-logang@deltatee.com> MIME-Version: 1.0 X-SA-Exim-Connect-IP: 172.16.1.31 X-SA-Exim-Rcpt-To: linux-kernel@vger.kernel.org, dmaengine@vger.kernel.org, vkoul@kernel.org, dan.j.williams@intel.com, dave.jiang@intel.com, kchow@gigaio.com, logang@deltatee.com X-SA-Exim-Mail-From: gunthorp@deltatee.com X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on ale.deltatee.com X-Spam-Level: X-Spam-Status: No, score=-6.5 required=5.0 tests=ALL_TRUSTED,BAYES_00, MYRULES_FREE,MYRULES_NO_TEXT autolearn=no autolearn_force=no version=3.4.2 Subject: [PATCH 2/5] dmaengine: Call module_put() after device_free_chan_resources() X-SA-Exim-Version: 4.2.1 (built Wed, 08 May 2019 21:11:16 +0000) X-SA-Exim-Scanned: Yes (on ale.deltatee.com) Sender: dmaengine-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: dmaengine@vger.kernel.org The module reference is taken to ensure the callbacks still exist when they are called. If the channel holds the last reference to the module, the module can disappear before device_free_chan_resources() is called and would cause a call into free'd memory. Signed-off-by: Logan Gunthorpe --- drivers/dma/dmaengine.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c index 4b604086b1b3..776fdf535a3a 100644 --- a/drivers/dma/dmaengine.c +++ b/drivers/dma/dmaengine.c @@ -250,7 +250,6 @@ static void dma_chan_put(struct dma_chan *chan) return; chan->client_count--; - module_put(dma_chan_to_owner(chan)); /* This channel is not in use anymore, free it */ if (!chan->client_count && chan->device->device_free_chan_resources) { @@ -259,6 +258,8 @@ static void dma_chan_put(struct dma_chan *chan) chan->device->device_free_chan_resources(chan); } + module_put(dma_chan_to_owner(chan)); + /* If the channel is used via a DMA request router, free the mapping */ if (chan->router && chan->router->route_free) { chan->router->route_free(chan->router->dev, chan->route_data); From patchwork Mon Dec 16 19:01:18 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Logan Gunthorpe X-Patchwork-Id: 11295059 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id D0A88112B for ; Mon, 16 Dec 2019 19:01:45 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id A3A55206E0 for ; Mon, 16 Dec 2019 19:01:45 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727140AbfLPTBl (ORCPT ); Mon, 16 Dec 2019 14:01:41 -0500 Received: from ale.deltatee.com ([207.54.116.67]:34952 "EHLO ale.deltatee.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726551AbfLPTB0 (ORCPT ); Mon, 16 Dec 2019 14:01:26 -0500 Received: from cgy1-donard.priv.deltatee.com ([172.16.1.31]) by ale.deltatee.com with esmtps (TLS1.3:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1igvcR-0005jF-3f; Mon, 16 Dec 2019 12:01:25 -0700 Received: from gunthorp by cgy1-donard.priv.deltatee.com with local (Exim 4.92) (envelope-from ) id 1igvcQ-0005Zg-GF; Mon, 16 Dec 2019 12:01:22 -0700 From: Logan Gunthorpe To: linux-kernel@vger.kernel.org, dmaengine@vger.kernel.org, Vinod Koul Cc: Dan Williams , Dave Jiang , Kit Chow , Logan Gunthorpe Date: Mon, 16 Dec 2019 12:01:18 -0700 Message-Id: <20191216190120.21374-4-logang@deltatee.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20191216190120.21374-1-logang@deltatee.com> References: <20191216190120.21374-1-logang@deltatee.com> MIME-Version: 1.0 X-SA-Exim-Connect-IP: 172.16.1.31 X-SA-Exim-Rcpt-To: linux-kernel@vger.kernel.org, dmaengine@vger.kernel.org, vkoul@kernel.org, dan.j.williams@intel.com, dave.jiang@intel.com, kchow@gigaio.com, logang@deltatee.com X-SA-Exim-Mail-From: gunthorp@deltatee.com X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on ale.deltatee.com X-Spam-Level: X-Spam-Status: No, score=-6.5 required=5.0 tests=ALL_TRUSTED,BAYES_00, MYRULES_FREE,MYRULES_NO_TEXT autolearn=no autolearn_force=no version=3.4.2 Subject: [PATCH 3/5] dmaengine: Move dma_channel_rebalance() infrastructure up in code X-SA-Exim-Version: 4.2.1 (built Wed, 08 May 2019 21:11:16 +0000) X-SA-Exim-Scanned: Yes (on ale.deltatee.com) Sender: dmaengine-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: dmaengine@vger.kernel.org So it can be called by a release function which is needed higher up in the code. No functional changes intended. Signed-off-by: Logan Gunthorpe --- drivers/dma/dmaengine.c | 288 ++++++++++++++++++++-------------------- 1 file changed, 144 insertions(+), 144 deletions(-) diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c index 776fdf535a3a..1f9a6293f15a 100644 --- a/drivers/dma/dmaengine.c +++ b/drivers/dma/dmaengine.c @@ -164,6 +164,150 @@ static struct class dma_devclass = { /* --- client and device registration --- */ +/** + * dma_cap_mask_all - enable iteration over all operation types + */ +static dma_cap_mask_t dma_cap_mask_all; + +/** + * dma_chan_tbl_ent - tracks channel allocations per core/operation + * @chan - associated channel for this entry + */ +struct dma_chan_tbl_ent { + struct dma_chan *chan; +}; + +/** + * channel_table - percpu lookup table for memory-to-memory offload providers + */ +static struct dma_chan_tbl_ent __percpu *channel_table[DMA_TX_TYPE_END]; + +static int __init dma_channel_table_init(void) +{ + enum dma_transaction_type cap; + int err = 0; + + bitmap_fill(dma_cap_mask_all.bits, DMA_TX_TYPE_END); + + /* 'interrupt', 'private', and 'slave' are channel capabilities, + * but are not associated with an operation so they do not need + * an entry in the channel_table + */ + clear_bit(DMA_INTERRUPT, dma_cap_mask_all.bits); + clear_bit(DMA_PRIVATE, dma_cap_mask_all.bits); + clear_bit(DMA_SLAVE, dma_cap_mask_all.bits); + + for_each_dma_cap_mask(cap, dma_cap_mask_all) { + channel_table[cap] = alloc_percpu(struct dma_chan_tbl_ent); + if (!channel_table[cap]) { + err = -ENOMEM; + break; + } + } + + if (err) { + pr_err("initialization failure\n"); + for_each_dma_cap_mask(cap, dma_cap_mask_all) + free_percpu(channel_table[cap]); + } + + return err; +} +arch_initcall(dma_channel_table_init); + +/** + * dma_chan_is_local - returns true if the channel is in the same numa-node as + * the cpu + */ +static bool dma_chan_is_local(struct dma_chan *chan, int cpu) +{ + int node = dev_to_node(chan->device->dev); + return node == NUMA_NO_NODE || + cpumask_test_cpu(cpu, cpumask_of_node(node)); +} + +/** + * min_chan - returns the channel with min count and in the same numa-node as + * the cpu + * @cap: capability to match + * @cpu: cpu index which the channel should be close to + * + * If some channels are close to the given cpu, the one with the lowest + * reference count is returned. Otherwise, cpu is ignored and only the + * reference count is taken into account. + * Must be called under dma_list_mutex. + */ +static struct dma_chan *min_chan(enum dma_transaction_type cap, int cpu) +{ + struct dma_device *device; + struct dma_chan *chan; + struct dma_chan *min = NULL; + struct dma_chan *localmin = NULL; + + list_for_each_entry(device, &dma_device_list, global_node) { + if (!dma_has_cap(cap, device->cap_mask) || + dma_has_cap(DMA_PRIVATE, device->cap_mask)) + continue; + list_for_each_entry(chan, &device->channels, device_node) { + if (!chan->client_count) + continue; + if (!min || chan->table_count < min->table_count) + min = chan; + + if (dma_chan_is_local(chan, cpu)) + if (!localmin || + chan->table_count < localmin->table_count) + localmin = chan; + } + } + + chan = localmin ? localmin : min; + + if (chan) + chan->table_count++; + + return chan; +} + +/** + * dma_channel_rebalance - redistribute the available channels + * + * Optimize for cpu isolation (each cpu gets a dedicated channel for an + * operation type) in the SMP case, and operation isolation (avoid + * multi-tasking channels) in the non-SMP case. Must be called under + * dma_list_mutex. + */ +static void dma_channel_rebalance(void) +{ + struct dma_chan *chan; + struct dma_device *device; + int cpu; + int cap; + + /* undo the last distribution */ + for_each_dma_cap_mask(cap, dma_cap_mask_all) + for_each_possible_cpu(cpu) + per_cpu_ptr(channel_table[cap], cpu)->chan = NULL; + + list_for_each_entry(device, &dma_device_list, global_node) { + if (dma_has_cap(DMA_PRIVATE, device->cap_mask)) + continue; + list_for_each_entry(chan, &device->channels, device_node) + chan->table_count = 0; + } + + /* don't populate the channel_table if no clients are available */ + if (!dmaengine_ref_count) + return; + + /* redistribute available channels */ + for_each_dma_cap_mask(cap, dma_cap_mask_all) + for_each_online_cpu(cpu) { + chan = min_chan(cap, cpu); + per_cpu_ptr(channel_table[cap], cpu)->chan = chan; + } +} + #define dma_device_satisfies_mask(device, mask) \ __dma_device_satisfies_mask((device), &(mask)) static int @@ -289,57 +433,6 @@ enum dma_status dma_sync_wait(struct dma_chan *chan, dma_cookie_t cookie) } EXPORT_SYMBOL(dma_sync_wait); -/** - * dma_cap_mask_all - enable iteration over all operation types - */ -static dma_cap_mask_t dma_cap_mask_all; - -/** - * dma_chan_tbl_ent - tracks channel allocations per core/operation - * @chan - associated channel for this entry - */ -struct dma_chan_tbl_ent { - struct dma_chan *chan; -}; - -/** - * channel_table - percpu lookup table for memory-to-memory offload providers - */ -static struct dma_chan_tbl_ent __percpu *channel_table[DMA_TX_TYPE_END]; - -static int __init dma_channel_table_init(void) -{ - enum dma_transaction_type cap; - int err = 0; - - bitmap_fill(dma_cap_mask_all.bits, DMA_TX_TYPE_END); - - /* 'interrupt', 'private', and 'slave' are channel capabilities, - * but are not associated with an operation so they do not need - * an entry in the channel_table - */ - clear_bit(DMA_INTERRUPT, dma_cap_mask_all.bits); - clear_bit(DMA_PRIVATE, dma_cap_mask_all.bits); - clear_bit(DMA_SLAVE, dma_cap_mask_all.bits); - - for_each_dma_cap_mask(cap, dma_cap_mask_all) { - channel_table[cap] = alloc_percpu(struct dma_chan_tbl_ent); - if (!channel_table[cap]) { - err = -ENOMEM; - break; - } - } - - if (err) { - pr_err("initialization failure\n"); - for_each_dma_cap_mask(cap, dma_cap_mask_all) - free_percpu(channel_table[cap]); - } - - return err; -} -arch_initcall(dma_channel_table_init); - /** * dma_find_channel - find a channel to carry out the operation * @tx_type: transaction type @@ -370,97 +463,6 @@ void dma_issue_pending_all(void) } EXPORT_SYMBOL(dma_issue_pending_all); -/** - * dma_chan_is_local - returns true if the channel is in the same numa-node as the cpu - */ -static bool dma_chan_is_local(struct dma_chan *chan, int cpu) -{ - int node = dev_to_node(chan->device->dev); - return node == NUMA_NO_NODE || - cpumask_test_cpu(cpu, cpumask_of_node(node)); -} - -/** - * min_chan - returns the channel with min count and in the same numa-node as the cpu - * @cap: capability to match - * @cpu: cpu index which the channel should be close to - * - * If some channels are close to the given cpu, the one with the lowest - * reference count is returned. Otherwise, cpu is ignored and only the - * reference count is taken into account. - * Must be called under dma_list_mutex. - */ -static struct dma_chan *min_chan(enum dma_transaction_type cap, int cpu) -{ - struct dma_device *device; - struct dma_chan *chan; - struct dma_chan *min = NULL; - struct dma_chan *localmin = NULL; - - list_for_each_entry(device, &dma_device_list, global_node) { - if (!dma_has_cap(cap, device->cap_mask) || - dma_has_cap(DMA_PRIVATE, device->cap_mask)) - continue; - list_for_each_entry(chan, &device->channels, device_node) { - if (!chan->client_count) - continue; - if (!min || chan->table_count < min->table_count) - min = chan; - - if (dma_chan_is_local(chan, cpu)) - if (!localmin || - chan->table_count < localmin->table_count) - localmin = chan; - } - } - - chan = localmin ? localmin : min; - - if (chan) - chan->table_count++; - - return chan; -} - -/** - * dma_channel_rebalance - redistribute the available channels - * - * Optimize for cpu isolation (each cpu gets a dedicated channel for an - * operation type) in the SMP case, and operation isolation (avoid - * multi-tasking channels) in the non-SMP case. Must be called under - * dma_list_mutex. - */ -static void dma_channel_rebalance(void) -{ - struct dma_chan *chan; - struct dma_device *device; - int cpu; - int cap; - - /* undo the last distribution */ - for_each_dma_cap_mask(cap, dma_cap_mask_all) - for_each_possible_cpu(cpu) - per_cpu_ptr(channel_table[cap], cpu)->chan = NULL; - - list_for_each_entry(device, &dma_device_list, global_node) { - if (dma_has_cap(DMA_PRIVATE, device->cap_mask)) - continue; - list_for_each_entry(chan, &device->channels, device_node) - chan->table_count = 0; - } - - /* don't populate the channel_table if no clients are available */ - if (!dmaengine_ref_count) - return; - - /* redistribute available channels */ - for_each_dma_cap_mask(cap, dma_cap_mask_all) - for_each_online_cpu(cpu) { - chan = min_chan(cap, cpu); - per_cpu_ptr(channel_table[cap], cpu)->chan = chan; - } -} - int dma_get_slave_caps(struct dma_chan *chan, struct dma_slave_caps *caps) { struct dma_device *device; @@ -1376,5 +1378,3 @@ static int __init dma_bus_init(void) return class_register(&dma_devclass); } arch_initcall(dma_bus_init); - - From patchwork Mon Dec 16 19:01:19 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Logan Gunthorpe X-Patchwork-Id: 11295057 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 9511014B7 for ; Mon, 16 Dec 2019 19:01:41 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 71EDC206E0 for ; Mon, 16 Dec 2019 19:01:41 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1725836AbfLPTBh (ORCPT ); Mon, 16 Dec 2019 14:01:37 -0500 Received: from ale.deltatee.com ([207.54.116.67]:34944 "EHLO ale.deltatee.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726467AbfLPTB0 (ORCPT ); Mon, 16 Dec 2019 14:01:26 -0500 Received: from cgy1-donard.priv.deltatee.com ([172.16.1.31]) by ale.deltatee.com with esmtps (TLS1.3:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1igvcR-0005jG-3f; Mon, 16 Dec 2019 12:01:24 -0700 Received: from gunthorp by cgy1-donard.priv.deltatee.com with local (Exim 4.92) (envelope-from ) id 1igvcQ-0005Zj-Jk; Mon, 16 Dec 2019 12:01:22 -0700 From: Logan Gunthorpe To: linux-kernel@vger.kernel.org, dmaengine@vger.kernel.org, Vinod Koul Cc: Dan Williams , Dave Jiang , Kit Chow , Logan Gunthorpe Date: Mon, 16 Dec 2019 12:01:19 -0700 Message-Id: <20191216190120.21374-5-logang@deltatee.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20191216190120.21374-1-logang@deltatee.com> References: <20191216190120.21374-1-logang@deltatee.com> MIME-Version: 1.0 X-SA-Exim-Connect-IP: 172.16.1.31 X-SA-Exim-Rcpt-To: linux-kernel@vger.kernel.org, dmaengine@vger.kernel.org, vkoul@kernel.org, dan.j.williams@intel.com, dave.jiang@intel.com, kchow@gigaio.com, logang@deltatee.com X-SA-Exim-Mail-From: gunthorp@deltatee.com X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on ale.deltatee.com X-Spam-Level: X-Spam-Status: No, score=-6.5 required=5.0 tests=ALL_TRUSTED,BAYES_00, MYRULES_FREE,MYRULES_NO_TEXT autolearn=no autolearn_force=no version=3.4.2 Subject: [PATCH 4/5] dmaengine: Add reference counting to dma_device struct X-SA-Exim-Version: 4.2.1 (built Wed, 08 May 2019 21:11:16 +0000) X-SA-Exim-Scanned: Yes (on ale.deltatee.com) Sender: dmaengine-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: dmaengine@vger.kernel.org Adding a reference count helps drivers to properly implement the unbind while in use case. References are taken and put every time a channel is allocated or freed. Once the final reference is put, the device is removed from the dma_device_list and a release callback function is called to signal the driver to free the memory. Signed-off-by: Logan Gunthorpe --- drivers/dma/dmaengine.c | 57 +++++++++++++++++++++++++++++++++------ include/linux/dmaengine.h | 8 +++++- 2 files changed, 56 insertions(+), 9 deletions(-) diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c index 1f9a6293f15a..e316abe3672d 100644 --- a/drivers/dma/dmaengine.c +++ b/drivers/dma/dmaengine.c @@ -342,6 +342,23 @@ static void balance_ref_count(struct dma_chan *chan) } } +static void dma_device_release(struct kref *ref) +{ + struct dma_device *device = container_of(ref, struct dma_device, ref); + + list_del_rcu(&device->global_node); + dma_channel_rebalance(); + + if (device->device_release) + device->device_release(device); +} + +static void dma_device_put(struct dma_device *device) +{ + lockdep_assert_held(&dma_list_mutex); + kref_put(&device->ref, dma_device_release); +} + /** * dma_chan_get - try to grab a dma channel's parent driver module * @chan - channel to grab @@ -362,6 +379,12 @@ static int dma_chan_get(struct dma_chan *chan) if (!try_module_get(owner)) return -ENODEV; + ret = kref_get_unless_zero(&chan->device->ref); + if (!ret) { + ret = -ENODEV; + goto module_put_out; + } + /* allocate upon first client reference */ if (chan->device->device_alloc_chan_resources) { ret = chan->device->device_alloc_chan_resources(chan); @@ -377,6 +400,8 @@ static int dma_chan_get(struct dma_chan *chan) return 0; err_out: + dma_device_put(chan->device); +module_put_out: module_put(owner); return ret; } @@ -402,6 +427,7 @@ static void dma_chan_put(struct dma_chan *chan) chan->device->device_free_chan_resources(chan); } + dma_device_put(chan->device); module_put(dma_chan_to_owner(chan)); /* If the channel is used via a DMA request router, free the mapping */ @@ -837,14 +863,14 @@ EXPORT_SYMBOL(dmaengine_get); */ void dmaengine_put(void) { - struct dma_device *device; + struct dma_device *device, *_d; struct dma_chan *chan; mutex_lock(&dma_list_mutex); dmaengine_ref_count--; BUG_ON(dmaengine_ref_count < 0); /* drop channel references */ - list_for_each_entry(device, &dma_device_list, global_node) { + list_for_each_entry_safe(device, _d, &dma_device_list, global_node) { if (dma_has_cap(DMA_PRIVATE, device->cap_mask)) continue; list_for_each_entry(chan, &device->channels, device_node) @@ -906,6 +932,10 @@ static int get_dma_id(struct dma_device *device) /** * dma_async_device_register - registers DMA devices found * @device: &dma_device + * + * After calling this routine the structure should not be freed except in the + * device_release() callback which will be called after + * dma_async_device_unregister() is called and no further references are taken. */ int dma_async_device_register(struct dma_device *device) { @@ -999,6 +1029,12 @@ int dma_async_device_register(struct dma_device *device) return -EIO; } + if (!device->device_release) + dev_warn(device->dev, + "WARN: Device release is not defined so it is not safe to unbind this driver while in use\n"); + + kref_init(&device->ref); + /* note: this only matters in the * CONFIG_ASYNC_TX_ENABLE_CHANNEL_SWITCH=n case */ @@ -1115,13 +1151,8 @@ void dma_async_device_unregister(struct dma_device *device) { struct dma_chan *chan; - mutex_lock(&dma_list_mutex); - list_del_rcu(&device->global_node); - dma_channel_rebalance(); - mutex_unlock(&dma_list_mutex); - list_for_each_entry(chan, &device->channels, device_node) { - WARN_ONCE(chan->client_count, + WARN_ONCE(!device->device_release && chan->client_count, "%s called while %d clients hold a reference\n", __func__, chan->client_count); mutex_lock(&dma_list_mutex); @@ -1130,6 +1161,16 @@ void dma_async_device_unregister(struct dma_device *device) device_unregister(&chan->dev->device); free_percpu(chan->local); } + + mutex_lock(&dma_list_mutex); + /* + * setting DMA_PRIVATE ensures the device being torn down will not + * be used in the channel_table + */ + dma_cap_set(DMA_PRIVATE, device->cap_mask); + dma_channel_rebalance(); + dma_device_put(device); + mutex_unlock(&dma_list_mutex); } EXPORT_SYMBOL(dma_async_device_unregister); diff --git a/include/linux/dmaengine.h b/include/linux/dmaengine.h index 13aa0abb71de..61e59a544af2 100644 --- a/include/linux/dmaengine.h +++ b/include/linux/dmaengine.h @@ -719,9 +719,14 @@ struct dma_filter { * will just return a simple status code * @device_issue_pending: push pending transactions to hardware * @descriptor_reuse: a submitted transfer can be resubmitted after completion + * @device_release: called sometime atfer dma_async_device_unregister() is + * called and there are no further references to this structure. This + * must be implemented to free resources however many existing drivers + * do not and are therefore not safe to unbind while in use. + * */ struct dma_device { - + struct kref ref; unsigned int chancnt; unsigned int privatecnt; struct list_head channels; @@ -802,6 +807,7 @@ struct dma_device { dma_cookie_t cookie, struct dma_tx_state *txstate); void (*device_issue_pending)(struct dma_chan *chan); + void (*device_release)(struct dma_device *dev); }; static inline int dmaengine_slave_config(struct dma_chan *chan, From patchwork Mon Dec 16 19:01:20 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Logan Gunthorpe X-Patchwork-Id: 11295063 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id D3D3714B7 for ; Mon, 16 Dec 2019 19:01:51 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id BB5D021739 for ; Mon, 16 Dec 2019 19:01:51 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726551AbfLPTBr (ORCPT ); Mon, 16 Dec 2019 14:01:47 -0500 Received: from ale.deltatee.com ([207.54.116.67]:34928 "EHLO ale.deltatee.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726390AbfLPTBZ (ORCPT ); Mon, 16 Dec 2019 14:01:25 -0500 Received: from cgy1-donard.priv.deltatee.com ([172.16.1.31]) by ale.deltatee.com with esmtps (TLS1.3:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1igvcR-0005jH-3f; Mon, 16 Dec 2019 12:01:23 -0700 Received: from gunthorp by cgy1-donard.priv.deltatee.com with local (Exim 4.92) (envelope-from ) id 1igvcQ-0005Zm-Mi; Mon, 16 Dec 2019 12:01:22 -0700 From: Logan Gunthorpe To: linux-kernel@vger.kernel.org, dmaengine@vger.kernel.org, Vinod Koul Cc: Dan Williams , Dave Jiang , Kit Chow , Logan Gunthorpe Date: Mon, 16 Dec 2019 12:01:20 -0700 Message-Id: <20191216190120.21374-6-logang@deltatee.com> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20191216190120.21374-1-logang@deltatee.com> References: <20191216190120.21374-1-logang@deltatee.com> MIME-Version: 1.0 X-SA-Exim-Connect-IP: 172.16.1.31 X-SA-Exim-Rcpt-To: linux-kernel@vger.kernel.org, dmaengine@vger.kernel.org, vkoul@kernel.org, dan.j.williams@intel.com, dave.jiang@intel.com, kchow@gigaio.com, logang@deltatee.com X-SA-Exim-Mail-From: gunthorp@deltatee.com X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on ale.deltatee.com X-Spam-Level: X-Spam-Status: No, score=-6.5 required=5.0 tests=ALL_TRUSTED,BAYES_00, MYRULES_FREE,MYRULES_NO_TEXT autolearn=no autolearn_force=no version=3.4.2 Subject: [PATCH 5/5] dmaengine: ioat: Support in-use unbind X-SA-Exim-Version: 4.2.1 (built Wed, 08 May 2019 21:11:16 +0000) X-SA-Exim-Scanned: Yes (on ale.deltatee.com) Sender: dmaengine-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: dmaengine@vger.kernel.org Don't allocate memory using the devm infrastructure and instead call kfree with the new dmaengine device_release call back. This ensures the structures are available until the last reference is dropped. We also need to ensure we call ioat_shutdown() in ioat_remove() so that all the channels are quiesced and further transaction fails. Signed-off-by: Logan Gunthorpe Acked-by: Dave Jiang --- drivers/dma/ioat/init.c | 38 ++++++++++++++++++++++++++------------ 1 file changed, 26 insertions(+), 12 deletions(-) diff --git a/drivers/dma/ioat/init.c b/drivers/dma/ioat/init.c index a6a6dc432db8..60e9afbb896c 100644 --- a/drivers/dma/ioat/init.c +++ b/drivers/dma/ioat/init.c @@ -556,10 +556,6 @@ static void ioat_dma_remove(struct ioatdma_device *ioat_dma) ioat_kobject_del(ioat_dma); dma_async_device_unregister(dma); - - dma_pool_destroy(ioat_dma->completion_pool); - - INIT_LIST_HEAD(&dma->channels); } /** @@ -589,7 +585,7 @@ static void ioat_enumerate_channels(struct ioatdma_device *ioat_dma) dev_dbg(dev, "%s: xfercap = %d\n", __func__, 1 << xfercap_log); for (i = 0; i < dma->chancnt; i++) { - ioat_chan = devm_kzalloc(dev, sizeof(*ioat_chan), GFP_KERNEL); + ioat_chan = kzalloc(sizeof(*ioat_chan), GFP_KERNEL); if (!ioat_chan) break; @@ -624,12 +620,16 @@ static void ioat_free_chan_resources(struct dma_chan *c) return; ioat_stop(ioat_chan); - ioat_reset_hw(ioat_chan); - /* Put LTR to idle */ - if (ioat_dma->version >= IOAT_VER_3_4) - writeb(IOAT_CHAN_LTR_SWSEL_IDLE, - ioat_chan->reg_base + IOAT_CHAN_LTR_SWSEL_OFFSET); + if (!test_bit(IOAT_CHAN_DOWN, &ioat_chan->state)) { + ioat_reset_hw(ioat_chan); + + /* Put LTR to idle */ + if (ioat_dma->version >= IOAT_VER_3_4) + writeb(IOAT_CHAN_LTR_SWSEL_IDLE, + ioat_chan->reg_base + + IOAT_CHAN_LTR_SWSEL_OFFSET); + } spin_lock_bh(&ioat_chan->cleanup_lock); spin_lock_bh(&ioat_chan->prep_lock); @@ -1322,16 +1322,28 @@ static struct pci_driver ioat_pci_driver = { .err_handler = &ioat_err_handler, }; +static void release_ioatdma(struct dma_device *device) +{ + struct ioatdma_device *d = to_ioatdma_device(device); + int i; + + for (i = 0; i < IOAT_MAX_CHANS; i++) + kfree(d->idx[i]); + + dma_pool_destroy(d->completion_pool); + kfree(d); +} + static struct ioatdma_device * alloc_ioatdma(struct pci_dev *pdev, void __iomem *iobase) { - struct device *dev = &pdev->dev; - struct ioatdma_device *d = devm_kzalloc(dev, sizeof(*d), GFP_KERNEL); + struct ioatdma_device *d = kzalloc(sizeof(*d), GFP_KERNEL); if (!d) return NULL; d->pdev = pdev; d->reg_base = iobase; + d->dma_dev.device_release = release_ioatdma; return d; } @@ -1400,6 +1412,8 @@ static void ioat_remove(struct pci_dev *pdev) if (!device) return; + ioat_shutdown(pdev); + dev_err(&pdev->dev, "Removing dma and dca services\n"); if (device->dca) { unregister_dca_provider(device->dca, &pdev->dev);