From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ozlabs.org (ozlabs.org [IPv6:2401:3900:2:1::2]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by lists.ozlabs.org (Postfix) with ESMTPS id 813381A0013 for ; Wed, 25 Feb 2015 16:13:45 +1100 (AEDT) Received: from e23smtp01.au.ibm.com (e23smtp01.au.ibm.com [202.81.31.143]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 397881400D5 for ; Wed, 25 Feb 2015 16:13:45 +1100 (AEDT) Received: from /spool/local by e23smtp01.au.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Wed, 25 Feb 2015 15:13:44 +1000 Received: from d23relay06.au.ibm.com (d23relay06.au.ibm.com [9.185.63.219]) by d23dlp02.au.ibm.com (Postfix) with ESMTP id 02E3F2BB004D for ; Wed, 25 Feb 2015 16:13:41 +1100 (EST) Received: from d23av03.au.ibm.com (d23av03.au.ibm.com [9.190.234.97]) by d23relay06.au.ibm.com (8.14.9/8.14.9/NCO v10.0) with ESMTP id t1P5DWsN40763580 for ; Wed, 25 Feb 2015 16:13:40 +1100 Received: from d23av03.au.ibm.com (localhost [127.0.0.1]) by d23av03.au.ibm.com (8.14.4/8.14.4/NCO v10.0 AVout) with ESMTP id t1P5D7rI025603 for ; Wed, 25 Feb 2015 16:13:07 +1100 From: "Ian Munsie" To: stable Subject: [PATCH] cxl: Fix issues when unmapping contexts Date: Wed, 25 Feb 2015 16:12:21 +1100 Message-Id: <1424841141-11783-1-git-send-email-imunsie@au.ibm.com> Cc: Michael Neuling , Ian Munsie , linuxppc-dev List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , From: Ian Munsie commit 0712dc7e73e59d79bcead5d5520acf4e9e917e87 upstream. for the 3.18 stable series An issue was introduced with "cxl: Unmap MMIO regions when detaching a context" (b123429e6a9e8d03aacf888d23262835f0081448) where closing a context normally could also unmap the problem state area of other contexts currently using the AFU. It was also discovered that after a context's MMIO space had been unmapped it would read 0s when accessing it, whereas the expected behaviour was for the access to fail altogether. In order to address these issues, this patch does two things: - Forced mmap unmapping is only done when we are forcefully detaching all contexts, and not in the normal detach path. Since the normal context close path is tied to the file release any mmaps must have already been released so we don't need to worry in that case. - The mmap path now uses a vm_operations_struct with a fault handler. The fault handler ensures that the context is in started state, otherwise it fails the access attempt with a SIGBUS. Fixes: b123429e6a9e ("cxl: Unmap MMIO regions when detaching a context") Signed-off-by: Ian Munsie Signed-off-by: Michael Ellerman --- drivers/misc/cxl/context.c | 82 +++++++++++++++++++++++++++++++++++----------- drivers/misc/cxl/file.c | 14 ++++---- 2 files changed, 71 insertions(+), 25 deletions(-) diff --git a/drivers/misc/cxl/context.c b/drivers/misc/cxl/context.c index 51fd6b5..d1b55fe 100644 --- a/drivers/misc/cxl/context.c +++ b/drivers/misc/cxl/context.c @@ -100,6 +100,46 @@ int cxl_context_init(struct cxl_context *ctx, struct cxl_afu *afu, bool master, return 0; } +static int cxl_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) +{ + struct cxl_context *ctx = vma->vm_file->private_data; + unsigned long address = (unsigned long)vmf->virtual_address; + u64 area, offset; + + offset = vmf->pgoff << PAGE_SHIFT; + + pr_devel("%s: pe: %i address: 0x%lx offset: 0x%llx\n", + __func__, ctx->pe, address, offset); + + if (ctx->afu->current_mode == CXL_MODE_DEDICATED) { + area = ctx->afu->psn_phys; + if (offset > ctx->afu->adapter->ps_size) + return VM_FAULT_SIGBUS; + } else { + area = ctx->psn_phys; + if (offset > ctx->psn_size) + return VM_FAULT_SIGBUS; + } + + mutex_lock(&ctx->status_mutex); + + if (ctx->status != STARTED) { + mutex_unlock(&ctx->status_mutex); + pr_devel("%s: Context not started, failing problem state access\n", __func__); + return VM_FAULT_SIGBUS; + } + + vm_insert_pfn(vma, address, (area + offset) >> PAGE_SHIFT); + + mutex_unlock(&ctx->status_mutex); + + return VM_FAULT_NOPAGE; +} + +static const struct vm_operations_struct cxl_mmap_vmops = { + .fault = cxl_mmap_fault, +}; + /* * Map a per-context mmio space into the given vma. */ @@ -108,26 +148,25 @@ int cxl_context_iomap(struct cxl_context *ctx, struct vm_area_struct *vma) u64 len = vma->vm_end - vma->vm_start; len = min(len, ctx->psn_size); - if (ctx->afu->current_mode == CXL_MODE_DEDICATED) { - vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); - return vm_iomap_memory(vma, ctx->afu->psn_phys, ctx->afu->adapter->ps_size); - } + if (ctx->afu->current_mode != CXL_MODE_DEDICATED) { + /* make sure there is a valid per process space for this AFU */ + if ((ctx->master && !ctx->afu->psa) || (!ctx->afu->pp_psa)) { + pr_devel("AFU doesn't support mmio space\n"); + return -EINVAL; + } - /* make sure there is a valid per process space for this AFU */ - if ((ctx->master && !ctx->afu->psa) || (!ctx->afu->pp_psa)) { - pr_devel("AFU doesn't support mmio space\n"); - return -EINVAL; + /* Can't mmap until the AFU is enabled */ + if (!ctx->afu->enabled) + return -EBUSY; } - /* Can't mmap until the AFU is enabled */ - if (!ctx->afu->enabled) - return -EBUSY; - pr_devel("%s: mmio physical: %llx pe: %i master:%i\n", __func__, ctx->psn_phys, ctx->pe , ctx->master); + vma->vm_flags |= VM_IO | VM_PFNMAP; vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); - return vm_iomap_memory(vma, ctx->psn_phys, len); + vma->vm_ops = &cxl_mmap_vmops; + return 0; } /* @@ -150,12 +189,6 @@ static void __detach_context(struct cxl_context *ctx) afu_release_irqs(ctx); flush_work(&ctx->fault_work); /* Only needed for dedicated process */ wake_up_all(&ctx->wq); - - /* Release Problem State Area mapping */ - mutex_lock(&ctx->mapping_lock); - if (ctx->mapping) - unmap_mapping_range(ctx->mapping, 0, 0, 1); - mutex_unlock(&ctx->mapping_lock); } /* @@ -184,6 +217,17 @@ void cxl_context_detach_all(struct cxl_afu *afu) * created and torn down after the IDR removed */ __detach_context(ctx); + + /* + * We are force detaching - remove any active PSA mappings so + * userspace cannot interfere with the card if it comes back. + * Easiest way to exercise this is to unbind and rebind the + * driver via sysfs while it is in use. + */ + mutex_lock(&ctx->mapping_lock); + if (ctx->mapping) + unmap_mapping_range(ctx->mapping, 0, 0, 1); + mutex_unlock(&ctx->mapping_lock); } mutex_unlock(&afu->contexts_lock); } diff --git a/drivers/misc/cxl/file.c b/drivers/misc/cxl/file.c index e9f2f10..b15d811 100644 --- a/drivers/misc/cxl/file.c +++ b/drivers/misc/cxl/file.c @@ -140,18 +140,20 @@ static long afu_ioctl_start_work(struct cxl_context *ctx, pr_devel("%s: pe: %i\n", __func__, ctx->pe); - mutex_lock(&ctx->status_mutex); - if (ctx->status != OPENED) { - rc = -EIO; - goto out; - } - + /* Do this outside the status_mutex to avoid a circular dependency with + * the locking in cxl_mmap_fault() */ if (copy_from_user(&work, uwork, sizeof(struct cxl_ioctl_start_work))) { rc = -EFAULT; goto out; } + mutex_lock(&ctx->status_mutex); + if (ctx->status != OPENED) { + rc = -EIO; + goto out; + } + /* * if any of the reserved fields are set or any of the unused * flags are set it's invalid -- 2.1.4