Linaro-mm-sig

linaro-mm-sig@lists.linaro.org

15 participants
3025 discussions

Re: [PATCH 0/2] Support direct I/O read and write for memory allocated by dmabuf

by Christian König

Am 10.07.24 um 16:35 schrieb Lei Liu: > > 在 2024/7/10 22:14, Christian König 写道: >> Am 10.07.24 um 15:57 schrieb Lei Liu: >>> Use vm_insert_page to establish a mapping for the memory allocated >>> by dmabuf, thus supporting direct I/O read and write; and fix the >>> issue of incorrect memory statistics after mapping dmabuf memory. >> >> Well big NAK to that! Direct I/O is intentionally disabled on DMA-bufs. > > Hello! Could you explain why direct_io is disabled on DMABUF? Is there > any historical reason for this? It's basically one of the most fundamental design decision of DMA-Buf. The attachment/map/fence model DMA-buf uses is not really compatible with direct I/O on the underlying pages. >> >> We already discussed enforcing that in the DMA-buf framework and this >> patch probably means that we should really do that. >> >> Regards, >> Christian. > > Thank you for your response. With the application of AI large model > edgeification, we urgently need support for direct_io on DMABUF to > read some very large files. Do you have any new solutions or plans for > this? We have seen similar projects over the years and all of those turned out to be complete shipwrecks. There is currently a patch set under discussion to give the network subsystem DMA-buf support. If you are interest in network direct I/O that could help. Additional to that a lot of GPU drivers support userptr usages, e.g. to import malloced memory into the GPU driver. You can then also do direct I/O on that malloced memory and the kernel will enforce correct handling with the GPU driver through MMU notifiers. But as far as I know a general DMA-buf based solution isn't possible. Regards, Christian. > > Regards, > Lei Liu. > >> >>> >>> Lei Liu (2): >>> mm: dmabuf_direct_io: Support direct_io for memory allocated by >>> dmabuf >>> mm: dmabuf_direct_io: Fix memory statistics error for dmabuf >>> allocated >>> memory with direct_io support >>> >>> drivers/dma-buf/heaps/system_heap.c | 5 +++-- >>> fs/proc/task_mmu.c | 8 +++++++- >>> include/linux/mm.h | 1 + >>> mm/memory.c | 15 ++++++++++----- >>> mm/rmap.c | 9 +++++---- >>> 5 files changed, 26 insertions(+), 12 deletions(-) >>> >>

1 year, 2 months

Re: [PATCH] kselftests: dmabuf-heaps: Ensure the driver name is null-terminated

by T.J. Mercier

On Mon, Jul 8, 2024 at 6:47 AM Zenghui Yu <yuzenghui(a)huawei.com> wrote: > > Even if a vgem device is configured in, we will skip the import_vgem_fd() > test almost every time. > > TAP version 13 > 1..11 > # Testing heap: system > # ======================================= > # Testing allocation and importing: > ok 1 # SKIP Could not open vgem -1 > > The problem is that we use the DRM_IOCTL_VERSION ioctl to query the driver > version information but leave the name field a non-null-terminated string. > Terminate it properly to actually test against the vgem device. Hm yeah. Looks like drm_copy_field resets version.name to the actual size of the name in the case of truncation, so maybe worth checking that too in case there is a name like "vgemfoo" that gets converted to "vgem\0" by this? > > Signed-off-by: Zenghui Yu <yuzenghui(a)huawei.com> > --- > tools/testing/selftests/dmabuf-heaps/dmabuf-heap.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/tools/testing/selftests/dmabuf-heaps/dmabuf-heap.c b/tools/testing/selftests/dmabuf-heaps/dmabuf-heap.c > index 5f541522364f..2fcc74998fa9 100644 > --- a/tools/testing/selftests/dmabuf-heaps/dmabuf-heap.c > +++ b/tools/testing/selftests/dmabuf-heaps/dmabuf-heap.c > @@ -32,6 +32,8 @@ static int check_vgem(int fd) > if (ret) > return 0; > > + name[4] = '\0'; > + > return !strcmp(name, "vgem"); > } > > -- > 2.33.0 >

1 year, 2 months

Re: [PATCH RFC 1/3] firmware: qcom: implement object invoke support

by Dmitry Baryshkov

On Thu, 4 Jul 2024 at 00:40, Amirreza Zarrabi <quic_azarrabi(a)quicinc.com> wrote: > > > > On 7/3/2024 10:13 PM, Dmitry Baryshkov wrote: > > On Tue, Jul 02, 2024 at 10:57:36PM GMT, Amirreza Zarrabi wrote: > >> Qualcomm TEE hosts Trusted Applications and Services that run in the > >> secure world. Access to these resources is provided using object > >> capabilities. A TEE client with access to the capability can invoke > >> the object and request a service. Similarly, TEE can request a service > >> from nonsecure world with object capabilities that are exported to secure > >> world. > >> > >> We provide qcom_tee_object which represents an object in both secure > >> and nonsecure world. TEE clients can invoke an instance of qcom_tee_object > >> to access TEE. TEE can issue a callback request to nonsecure world > >> by invoking an instance of qcom_tee_object in nonsecure world. > > > > Please see Documentation/process/submitting-patches.rst on how to write > > commit messages. > > Ack. > > > > >> > >> Any driver in nonsecure world that is interested to export a struct (or a > >> service object) to TEE, requires to embed an instance of qcom_tee_object in > >> the relevant struct and implements the dispatcher function which is called > >> when TEE invoked the service object. > >> > >> We also provids simplified API which implements the Qualcomm TEE transport > >> protocol. The implementation is independent from any services that may > >> reside in nonsecure world. > > > > "also" usually means that it should go to a separate commit. > > I will split this patch to multiple smaller ones. > [...] > > > >> + } in, out; > >> +}; > >> + > >> +int qcom_tee_object_do_invoke(struct qcom_tee_object_invoke_ctx *oic, > >> + struct qcom_tee_object *object, unsigned long op, struct qcom_tee_arg u[], int *result); > > > > What's the difference between a result that gets returned by the > > function and the result that gets retuned via the pointer? > > The function result, is local to kernel, for instance memory allocation failure, > or failure to issue the smc call. The result in pointer, is the remote result, > for instance return value from TA, or the TEE itself. > > I'll use better name, e.g. 'remote_result'? See how this is handled by other parties. For example, PSCI. If you have a standard set of return codes, translate them to -ESOMETHING in your framework and let everybody else see only the standard errors. -- With best wishes Dmitry

1 year, 2 months

Re: [PATCH] dma-buf: Remove unnecessary kmalloc() cast

by Christoph Hellwig

On Mon, Jul 01, 2024 at 11:26:34PM -0700, Andrew Morton wrote: > No, I do think the cast is useful: > > struct page *page = dma_fence_chain_alloc(); > > will presently generate a warning. We want this. Your change will > remove that useful warning. > > > Unrelatedly: there is no earthly reason why this is implemented as a > macro. A static inline function would be so much better. Why do we > keep doing this. Agreed with all of the above. Adding the dmabuf maintainers.

1 year, 2 months

Re: [PATCH RFC 1/3] firmware: qcom: implement object invoke support

by Dmitry Baryshkov

On Tue, Jul 02, 2024 at 10:57:36PM GMT, Amirreza Zarrabi wrote: > Qualcomm TEE hosts Trusted Applications and Services that run in the > secure world. Access to these resources is provided using object > capabilities. A TEE client with access to the capability can invoke > the object and request a service. Similarly, TEE can request a service > from nonsecure world with object capabilities that are exported to secure > world. > > We provide qcom_tee_object which represents an object in both secure > and nonsecure world. TEE clients can invoke an instance of qcom_tee_object > to access TEE. TEE can issue a callback request to nonsecure world > by invoking an instance of qcom_tee_object in nonsecure world. Please see Documentation/process/submitting-patches.rst on how to write commit messages. > > Any driver in nonsecure world that is interested to export a struct (or a > service object) to TEE, requires to embed an instance of qcom_tee_object in > the relevant struct and implements the dispatcher function which is called > when TEE invoked the service object. > > We also provids simplified API which implements the Qualcomm TEE transport > protocol. The implementation is independent from any services that may > reside in nonsecure world. "also" usually means that it should go to a separate commit. > > Signed-off-by: Amirreza Zarrabi <quic_azarrabi(a)quicinc.com> > --- > drivers/firmware/qcom/Kconfig | 14 + > drivers/firmware/qcom/Makefile | 2 + > drivers/firmware/qcom/qcom_object_invoke/Makefile | 4 + > drivers/firmware/qcom/qcom_object_invoke/async.c | 142 +++ > drivers/firmware/qcom/qcom_object_invoke/core.c | 1139 ++++++++++++++++++++ > drivers/firmware/qcom/qcom_object_invoke/core.h | 186 ++++ > .../qcom/qcom_object_invoke/qcom_scm_invoke.c | 22 + > .../firmware/qcom/qcom_object_invoke/release_wq.c | 90 ++ > include/linux/firmware/qcom/qcom_object_invoke.h | 233 ++++ > 9 files changed, 1832 insertions(+) > > diff --git a/drivers/firmware/qcom/Kconfig b/drivers/firmware/qcom/Kconfig > index 7f6eb4174734..103ab82bae9f 100644 > --- a/drivers/firmware/qcom/Kconfig > +++ b/drivers/firmware/qcom/Kconfig > @@ -84,4 +84,18 @@ config QCOM_QSEECOM_UEFISECAPP > Select Y here to provide access to EFI variables on the aforementioned > platforms. > > +config QCOM_OBJECT_INVOKE_CORE > + bool "Secure TEE Communication Support" tristate > + help > + Various Qualcomm SoCs have a Trusted Execution Environment (TEE) running > + in the Trust Zone. This module provides an interface to that via the > + capability based object invocation, using SMC calls. > + > + OBJECT_INVOKE_CORE allows capability based secure communication between > + TEE and VMs. Using OBJECT_INVOKE_CORE, kernel can issue calls to TEE or > + TAs to request a service or exposes services to TEE and TAs. It implements > + the necessary marshaling of messages with TEE. > + > + Select Y here to provide access to TEE. > + > endmenu > diff --git a/drivers/firmware/qcom/Makefile b/drivers/firmware/qcom/Makefile > index 0be40a1abc13..dd5e00215b2e 100644 > --- a/drivers/firmware/qcom/Makefile > +++ b/drivers/firmware/qcom/Makefile > @@ -8,3 +8,5 @@ qcom-scm-objs += qcom_scm.o qcom_scm-smc.o qcom_scm-legacy.o > obj-$(CONFIG_QCOM_TZMEM) += qcom_tzmem.o > obj-$(CONFIG_QCOM_QSEECOM) += qcom_qseecom.o > obj-$(CONFIG_QCOM_QSEECOM_UEFISECAPP) += qcom_qseecom_uefisecapp.o > + > +obj-y += qcom_object_invoke/ > diff --git a/drivers/firmware/qcom/qcom_object_invoke/Makefile b/drivers/firmware/qcom/qcom_object_invoke/Makefile > new file mode 100644 > index 000000000000..6ef4d54891a5 > --- /dev/null > +++ b/drivers/firmware/qcom/qcom_object_invoke/Makefile > @@ -0,0 +1,4 @@ > +# SPDX-License-Identifier: GPL-2.0-only > + > +obj-$(CONFIG_QCOM_OBJECT_INVOKE_CORE) += object-invoke-core.o > +object-invoke-core-objs := qcom_scm_invoke.o release_wq.o async.o core.o > diff --git a/drivers/firmware/qcom/qcom_object_invoke/async.c b/drivers/firmware/qcom/qcom_object_invoke/async.c > new file mode 100644 > index 000000000000..dd022ec68d8b > --- /dev/null > +++ b/drivers/firmware/qcom/qcom_object_invoke/async.c > @@ -0,0 +1,142 @@ > +// SPDX-License-Identifier: GPL-2.0-only > +/* > + * Copyright (c) 2023 Qualcomm Innovation Center, Inc. All rights reserved. > + */ > + > +#include <linux/kobject.h> > +#include <linux/slab.h> > +#include <linux/mutex.h> > + > +#include "core.h" > + > +/* Async handlers and providers. */ > +struct async_msg { > + struct { > + u32 version; /* Protocol version: top 16b major, lower 16b minor. */ > + u32 op; /* Async operation. */ > + } header; > + > + /* Format of the Async data field is defined by the specified operation. */ > + > + struct { > + u32 count; /* Number of objects that should be released. */ > + u32 obj[]; > + } op_release; > +}; Another generic comment: please select some prefix (like QTEE_ / qtee_) and use it for _all_ defines and all names in the driver. `struct async_msg` means that it is some genric code that is applicable to the whole kernel. > + > +/* Async Operations and header information. */ > + > +#define ASYNC_HEADER_SIZE sizeof(((struct async_msg *)(0))->header) Extract struct definition. Use sizeof(struct qtee_async_msg_header). > + > +/* ASYNC_OP_x: operation. > + * ASYNC_OP_x_HDR_SIZE: header size for the operation. > + * ASYNC_OP_x_SIZE: size of each entry in a message for the operation. > + * ASYNC_OP_x_MSG_SIZE: size of a message with n entries. > + */ > + > +#define ASYNC_OP_RELEASE QCOM_TEE_OBJECT_OP_RELEASE /* Added in minor version 0x0000. **/ Anything before minor version 0x0000 ? > +#define ASYNC_OP_RELEASE_HDR_SIZE offsetof(struct async_msg, op_release.obj) > +#define ASYNC_OP_RELEASE_SIZE sizeof(((struct async_msg *)(0))->op_release.obj[0]) sizeof(u32) is much better > +#define ASYNC_OP_RELEASE_MSG_SIZE(n) \ > + (ASYNC_OP_RELEASE_HDR_SIZE + ((n) * ASYNC_OP_RELEASE_SIZE)) struct_size(). But I think you should be able to inline and/or drop most of these defines. > + > +/* async_qcom_tee_buffer return the available async buffer in the output buffer. */ > + > +static struct qcom_tee_buffer async_qcom_tee_buffer(struct qcom_tee_object_invoke_ctx *oic) Why do you need to return struct instance? > +{ > + int i; > + size_t offset; > + > + struct qcom_tee_callback *msg = (struct qcom_tee_callback *)oic->out.msg.addr; > + > + if (!(oic->flags & OIC_FLAG_BUSY)) > + return oic->out.msg; > + > + /* Async requests are appended to the output buffer after the CB message. */ > + > + offset = OFFSET_TO_BUFFER_ARGS(msg, counts_total(msg->counts)); > + > + for_each_input_buffer(i, msg->counts) > + offset += align_offset(msg->args[i].b.size); > + > + for_each_output_buffer(i, msg->counts) > + offset += align_offset(msg->args[i].b.size); > + > + if (oic->out.msg.size > offset) { > + return (struct qcom_tee_buffer) > + { { oic->out.msg.addr + offset }, oic->out.msg.size - offset }; > + } > + > + pr_err("no space left for async messages! or malformed message.\n"); No spamming on the kmsg. > + > + return (struct qcom_tee_buffer) { { 0 }, 0 }; This doesn't look correct. > +} > + What does this function return? > +static size_t async_release_handler(struct qcom_tee_object_invoke_ctx *oic, > + struct async_msg *async_msg, size_t size) Please ident the code properly, this should be aligned to the open bracket. > +{ > + int i; > + > + /* We need space for at least a single entry. */ > + if (size < ASYNC_OP_RELEASE_MSG_SIZE(1)) > + return 0; > + > + for (i = 0; i < async_msg->op_release.count; i++) { > + struct qcom_tee_object *object; > + > + /* Remove the object from xa_qcom_tee_objects so that the object_id > + * becomes invalid for further use. However, call put_qcom_tee_object > + * to schedule the actual release if there is no user. > + */ > + > + object = erase_qcom_tee_object(async_msg->op_release.obj[i]); > + > + put_qcom_tee_object(object); > + } > + > + return ASYNC_OP_RELEASE_MSG_SIZE(i); > +} > + > +/* '__fetch__async_reqs' is a handler dispatcher (from TEE). */ > + > +void __fetch__async_reqs(struct qcom_tee_object_invoke_ctx *oic) > +{ > + size_t consumed, used = 0; > + > + struct qcom_tee_buffer async_buffer = async_qcom_tee_buffer(oic); > + > + while (async_buffer.size - used > ASYNC_HEADER_SIZE) { > + struct async_msg *async_msg = (struct async_msg *)(async_buffer.addr + used); > + > + /* TEE assumes unused buffer is set to zero. */ > + if (!async_msg->header.version) > + goto out; > + > + switch (async_msg->header.op) { > + case ASYNC_OP_RELEASE: > + consumed = async_release_handler(oic, > + async_msg, async_buffer.size - used); > + > + break; > + default: /* Unsupported operations. */ > + consumed = 0; > + } > + > + used += align_offset(consumed); > + > + if (!consumed) { > + pr_err("Drop async buffer (context_id %d): buffer %p, (%p, %zx), processed %zx\n", Should it really go to the kmsg? > + oic->context_id, > + oic->out.msg.addr, /* Address of Output buffer. */ > + async_buffer.addr, /* Address of beginning of async buffer. */ > + async_buffer.size, /* Available size of async buffer. */ > + used); /* Processed async buffer. */ > + > + goto out; > + } > + } > + > + out: > + > + memset(async_buffer.addr, 0, async_buffer.size); Why? > +} > diff --git a/drivers/firmware/qcom/qcom_object_invoke/core.c b/drivers/firmware/qcom/qcom_object_invoke/core.c > new file mode 100644 > index 000000000000..37dde8946b08 > --- /dev/null > +++ b/drivers/firmware/qcom/qcom_object_invoke/core.c > @@ -0,0 +1,1139 @@ > +// SPDX-License-Identifier: GPL-2.0-only > +/* > + * Copyright (c) 2024 Qualcomm Innovation Center, Inc. All rights reserved. > + */ > + > +#include <linux/kobject.h> > +#include <linux/sysfs.h> > +#include <linux/module.h> > +#include <linux/init.h> > +#include <linux/slab.h> > +#include <linux/delay.h> > +#include <linux/mm.h> > +#include <linux/xarray.h> > + > +#include "core.h" > + > +/* Static 'Primordial Object' operations. */ > + > +#define OBJECT_OP_YIELD 1 > +#define OBJECT_OP_SLEEP 2 > + > +/* static_qcom_tee_object_primordial always exists. */ > +/* primordial_object_register and primordial_object_release extends it. */ > + > +static struct qcom_tee_object static_qcom_tee_object_primordial; > + > +static int primordial_object_register(struct qcom_tee_object *object); > +static void primordial_object_release(struct qcom_tee_object *object); > + > +/* Marshaling API. */ > +/* > + * prepare_msg - Prepares input buffer for sending to TEE. > + * update_args - Parses TEE response in input buffer. > + * prepare_args - Parses TEE request from output buffer. > + * update_msg - Updates output buffer with response for TEE request. > + * > + * prepare_msg and update_args are used in direct TEE object invocation. > + * prepare_args and update_msg are used for TEE requests (callback or async). > + */ > + > +static int prepare_msg(struct qcom_tee_object_invoke_ctx *oic, > + struct qcom_tee_object *object, unsigned long op, struct qcom_tee_arg u[]); > +static int update_args(struct qcom_tee_arg u[], struct qcom_tee_object_invoke_ctx *oic); > +static int prepare_args(struct qcom_tee_object_invoke_ctx *oic); > +static int update_msg(struct qcom_tee_object_invoke_ctx *oic); Please reorder the functions so that you don't need forward declarations. > + > +static int next_arg_type(struct qcom_tee_arg u[], int i, enum qcom_tee_arg_type type) > +{ > + while (u[i].type != QCOM_TEE_ARG_TYPE_END && u[i].type != type) > + i++; > + > + return i; > +} > + > +/** > + * args_for_each_type - Iterate over argument of given type. > + * @i: index in @args. > + * @args: array of arguments. > + * @at: type of argument. > + */ > +#define args_for_each_type(i, args, at) \ > + for (i = 0, i = next_arg_type(args, i, at); \ > + args[i].type != QCOM_TEE_ARG_TYPE_END; i = next_arg_type(args, ++i, at)) > + > +#define arg_for_each_input_buffer(i, args) args_for_each_type(i, args, QCOM_TEE_ARG_TYPE_IB) > +#define arg_for_each_output_buffer(i, args) args_for_each_type(i, args, QCOM_TEE_ARG_TYPE_OB) > +#define arg_for_each_input_object(i, args) args_for_each_type(i, args, QCOM_TEE_ARG_TYPE_IO) > +#define arg_for_each_output_object(i, args) args_for_each_type(i, args, QCOM_TEE_ARG_TYPE_OO) > + > +/* Outside this file we use struct qcom_tee_object to identify an object. */ > + > +/* We only allocate IDs with QCOM_TEE_OBJ_NS_BIT set in range > + * [QCOM_TEE_OBJECT_ID_START .. QCOM_TEE_OBJECT_ID_END]. qcom_tee_object > + * represents non-secure object. The first ID with QCOM_TEE_OBJ_NS_BIT set is reserved > + * for primordial object. > + */ > + > +#define QCOM_TEE_OBJECT_PRIMORDIAL (QCOM_TEE_OBJ_NS_BIT) > +#define QCOM_TEE_OBJECT_ID_START (QCOM_TEE_OBJECT_PRIMORDIAL + 1) > +#define QCOM_TEE_OBJECT_ID_END (UINT_MAX) > + > +#define SET_QCOM_TEE_OBJECT(p, type, ...) __SET_QCOM_TEE_OBJECT(p, type, ##__VA_ARGS__, 0UL) > +#define __SET_QCOM_TEE_OBJECT(p, type, optr, ...) do { \ > + (p)->object_type = (type); \ > + (p)->info.object_ptr = (unsigned long)(optr); \ > + (p)->release = NULL; \ > + } while (0) > + > +/* ''TEE Object Table''. */ > +static DEFINE_XARRAY_ALLOC(xa_qcom_tee_objects); > + > +struct qcom_tee_object *allocate_qcom_tee_object(void) > +{ > + struct qcom_tee_object *object; I thought that struct qcom_tee_object should be embedded into other struct definitions. Here you are just allocing it. Did I misunderstand something? > + > + object = kzalloc(sizeof(*object), GFP_KERNEL); > + if (object) > + SET_QCOM_TEE_OBJECT(object, QCOM_TEE_OBJECT_TYPE_NULL); > + > + return object; > +} > +EXPORT_SYMBOL_GPL(allocate_qcom_tee_object); kerneldoc for all exported functions. > + > +void free_qcom_tee_object(struct qcom_tee_object *object) > +{ > + kfree(object); > +} > +EXPORT_SYMBOL_GPL(free_qcom_tee_object); If qcom_tee_object is refcounted, then such API is defintely forbidden. > + > +/* 'get_qcom_tee_object' and 'put_qcom_tee_object'. */ > + > +static int __free_qcom_tee_object(struct qcom_tee_object *object); What is the difference between free_qcom_tee_object() and __free_qcom_tee_object() ? > +static void ____destroy_qcom_tee_object(struct kref *refcount) > +{ > + struct qcom_tee_object *object = container_of(refcount, struct qcom_tee_object, refcount); > + > + __free_qcom_tee_object(object); > +} > + > +int get_qcom_tee_object(struct qcom_tee_object *object) > +{ > + if (object != NULL_QCOM_TEE_OBJECT && > + object != ROOT_QCOM_TEE_OBJECT) > + return kref_get_unless_zero(&object->refcount); > + > + return 0; > +} > +EXPORT_SYMBOL_GPL(get_qcom_tee_object); > + > +static struct qcom_tee_object *qcom_tee__get_qcom_tee_object(unsigned int object_id) > +{ > + XA_STATE(xas, &xa_qcom_tee_objects, object_id); > + struct qcom_tee_object *object; > + > + rcu_read_lock(); > + do { > + object = xas_load(&xas); > + if (xa_is_zero(object)) > + object = NULL_QCOM_TEE_OBJECT; > + > + } while (xas_retry(&xas, object)); If you are just looping over the objects, why do you need XArray instead of list? > + > + /* Sure object still exists. */ Why? > + if (!get_qcom_tee_object(object)) > + object = NULL_QCOM_TEE_OBJECT; > + > + rcu_read_unlock(); > + > + return object; > +} > + > +struct qcom_tee_object *qcom_tee_get_qcom_tee_object(unsigned int object_id) > +{ > + switch (object_id) { > + case QCOM_TEE_OBJECT_PRIMORDIAL: > + return &static_qcom_tee_object_primordial; > + > + default: > + return qcom_tee__get_qcom_tee_object(object_id); > + } > +} > + > +void put_qcom_tee_object(struct qcom_tee_object *object) > +{ > + if (object != &static_qcom_tee_object_primordial && > + object != NULL_QCOM_TEE_OBJECT && > + object != ROOT_QCOM_TEE_OBJECT) misaligned > + kref_put(&object->refcount, ____destroy_qcom_tee_object); > +} > +EXPORT_SYMBOL_GPL(put_qcom_tee_object); > + > +/* 'alloc_qcom_tee_object_id' and 'erase_qcom_tee_object'. */ huh? I think I'm going to stop here. Please: - Split the driver into logical chunks and more logical pieces. - Rename functions and structures to follow generic scheme. Usually it is prefix_object_do_something(). - Add documentation that would allow us to understand what is going on. - Also document some general design decisions. Usage of XArray. API choice. Refcounting. [skipped] > diff --git a/include/linux/firmware/qcom/qcom_object_invoke.h b/include/linux/firmware/qcom/qcom_object_invoke.h > new file mode 100644 > index 000000000000..9e6acd0f4db0 > --- /dev/null > +++ b/include/linux/firmware/qcom/qcom_object_invoke.h > @@ -0,0 +1,233 @@ > +/* SPDX-License-Identifier: GPL-2.0-only */ > +/* > + * Copyright (c) 2023 Qualcomm Innovation Center, Inc. All rights reserved. > + */ > + > +#ifndef __QCOM_OBJECT_INVOKE_H > +#define __QCOM_OBJECT_INVOKE_H > + > +#include <linux/kref.h> > +#include <linux/completion.h> > +#include <linux/workqueue.h> > +#include <uapi/misc/qcom_tee.h> This header doesn't exist yet. This obviously means that you haven't actually tried building this patch. Please make sure that kernel compiles successfully after each commit. > + > +struct qcom_tee_object; > + > +/* Primordial Object */ > + > +/* It is used for bootstrapping the IPC connection between a VM and TEE. > + * > + * Each side (both the VM and the TEE) starts up with no object received from the > + * other side. They both ''assume'' the other side implements a permanent initial > + * object in the object table. > + * > + * TEE's initial object is typically called the ''root client env'', and it's > + * invoked by VMs when they want to get a new clientEnv. The initial object created > + * by the VMs is invoked by TEE, it's typically called the ''primordial object''. > + * > + * VM can register a primordial object using 'init_qcom_tee_object_user' with > + * 'QCOM_TEE_OBJECT_TYPE_ROOT' type. > + */ > + > +enum qcom_tee_object_type { > + QCOM_TEE_OBJECT_TYPE_USER = 0x1, /* TEE object. */ > + QCOM_TEE_OBJECT_TYPE_CB_OBJECT = 0x2, /* Callback Object. */ > + QCOM_TEE_OBJECT_TYPE_ROOT = 0x8, /* ''Root client env.'' or 'primordial' Object. */ > + QCOM_TEE_OBJECT_TYPE_NULL = 0x10, /* NULL object. */ > +}; > + > +enum qcom_tee_arg_type { > + QCOM_TEE_ARG_TYPE_END = 0, > + QCOM_TEE_ARG_TYPE_IB = 0x80, /* Input Buffer (IB). */ > + QCOM_TEE_ARG_TYPE_OB = 0x1, /* Output Buffer (OB). */ > + QCOM_TEE_ARG_TYPE_IO = 0x81, /* Input Object (IO). */ > + QCOM_TEE_ARG_TYPE_OO = 0x2 /* Output Object (OO). */ > +}; > + > +#define QCOM_TEE_ARG_TYPE_INPUT_MASK 0x80 > + > +/* Maximum specific type of arguments (i.e. IB, OB, IO, and OO) that can fit in a TEE message. */ > +#define QCOM_TEE_ARGS_PER_TYPE 16 Why is it 16? > + > +/* Maximum arguments that can fit in a TEE message. */ > +#define QCOM_TEE_ARGS_MAX (QCOM_TEE_ARGS_PER_TYPE * 4) > + > +/** > + * struct qcom_tee_arg - Argument for TEE object invocation. > + * @type: type of argument > + * @flags: extra flags. > + * @b: address and size if type of argument is buffer. > + * @o: qcom_tee_object instance if type of argument is object. > + * > + * @flags only accept QCOM_TEE_ARG_FLAGS_UADDR for now which states that @b > + * contains userspace address in uaddr. > + * > + */ > +struct qcom_tee_arg { > + enum qcom_tee_arg_type type; > + > +/* 'uaddr' holds a __user address. */ > +#define QCOM_TEE_ARG_FLAGS_UADDR 1 > + char flags; This is not a character. > + union { > + struct qcom_tee_buffer { > + union { > + void *addr; > + void __user *uaddr; > + }; > + size_t size; > + } b; > + struct qcom_tee_object *o; > + }; How can the code distinguish between the qcom_tee_object and qcom_tee_buffer here? > +}; > + > +static inline int size_of_arg(struct qcom_tee_arg u[]) length, not size. > +{ > + int i = 0; > + > + while (u[i].type != QCOM_TEE_ARG_TYPE_END) > + i++; > + > + return i; > +} > + > +/* Context ID - It is a unique ID assigned to a invocation which is in progress. > + * Objects's dispatcher can use the ID to differentiate between concurrent calls. > + * ID [0 .. 10) are reserved, i.e. never passed to object's dispatcher. Is 10 included or excluded here? Why does it have a different bracket type? > + */ > + > +struct qcom_tee_object_invoke_ctx { > + unsigned int context_id; > + > +#define OIC_FLAG_BUSY 1 /* Context is busy, i.e. callback is in progress. */ > +#define OIC_FLAG_NOTIFY 2 /* Context needs to notify the current object. */ > +#define OIC_FLAG_QCOM_TEE 4 /* Context has objects shared with TEE. */ BIT(n) > + unsigned int flags; > + > + /* Current object invoked in this callback context. */ > + struct qcom_tee_object *object; > + > + /* Arguments passed to dispatch callback. */ > + struct qcom_tee_arg u[QCOM_TEE_ARGS_MAX + 1]; > + > + int errno; > + > + /* Inbound and Outbound buffers shared with TEE. */ > + struct { > + struct qcom_tee_buffer msg; Please define struct qcom_tee_buffer in a top-level definition instead of having it nested somewhere in another struct; > + } in, out; > +}; > + > +int qcom_tee_object_do_invoke(struct qcom_tee_object_invoke_ctx *oic, > + struct qcom_tee_object *object, unsigned long op, struct qcom_tee_arg u[], int *result); What's the difference between a result that gets returned by the function and the result that gets retuned via the pointer? > + > +#define QCOM_TEE_OBJECT_OP_METHOD_MASK 0x0000FFFFU > +#define QCOM_TEE_OBJECT_OP_METHOD_ID(op) ((op) & QCOM_TEE_OBJECT_OP_METHOD_MASK) > + > +/* Reserved Operations. */ > + > +#define QCOM_TEE_OBJECT_OP_RELEASE (QCOM_TEE_OBJECT_OP_METHOD_MASK - 0) > +#define QCOM_TEE_OBJECT_OP_RETAIN (QCOM_TEE_OBJECT_OP_METHOD_MASK - 1) > +#define QCOM_TEE_OBJECT_OP_NO_OP (QCOM_TEE_OBJECT_OP_METHOD_MASK - 2) > + > +struct qcom_tee_object_operations { > + void (*release)(struct qcom_tee_object *object); > + > + /** > + * @op_supported: > + * > + * Query made to make sure the requested operation is supported. If defined, > + * it is called before marshaling of the arguments (as optimisation). > + */ > + int (*op_supported)(unsigned long op); > + > + /** > + * @notify: > + * > + * After @dispatch returned, it is called to notify the status of the transport; > + * i.e. transport errors or success. This allows the client to cleanup, if > + * the transport fails after @dispatch submits a SUCCESS response. > + */ > + void (*notify)(unsigned int context_id, struct qcom_tee_object *object, int status); > + > + int (*dispatch)(unsigned int context_id, struct qcom_tee_object *object, > + unsigned long op, struct qcom_tee_arg args[]); > + > + /** > + * @param_to_object: > + * > + * Called by core to do the object dependent marshaling from @param to an > + * instance of @object (NOT IMPLEMENTED YET). > + */ > + int (*param_to_object)(struct qcom_tee_param *param, struct qcom_tee_object *object); > + > + int (*object_to_param)(struct qcom_tee_object *object, struct qcom_tee_param *param); > +}; > + > +struct qcom_tee_object { > + const char *name; > + struct kref refcount; > + > + enum qcom_tee_object_type object_type; > + union object_info { > + unsigned long object_ptr; > + } info; > + > + struct qcom_tee_object_operations *ops; > + > + /* see release_wq.c. */ > + struct work_struct work; > + > + /* Callback for any internal cleanup before the object's release. */ > + void (*release)(struct qcom_tee_object *object); > +}; > + > +/* Static instances of qcom_tee_object objects. */ > + > +#define NULL_QCOM_TEE_OBJECT ((struct qcom_tee_object *)(0)) > + > +/* ROOT_QCOM_TEE_OBJECT aka ''root client env''. */ > +#define ROOT_QCOM_TEE_OBJECT ((struct qcom_tee_object *)(1)) My gut feeling is that an invalid non-null pointer is a path to disaster. > + > +static inline enum qcom_tee_object_type typeof_qcom_tee_object(struct qcom_tee_object *object) > +{ > + if (object == NULL_QCOM_TEE_OBJECT) > + return QCOM_TEE_OBJECT_TYPE_NULL; > + > + if (object == ROOT_QCOM_TEE_OBJECT) > + return QCOM_TEE_OBJECT_TYPE_ROOT; > + > + return object->object_type; > +} > + > +static inline const char *qcom_tee_object_name(struct qcom_tee_object *object) > +{ > + if (object == NULL_QCOM_TEE_OBJECT) > + return "null"; > + > + if (object == ROOT_QCOM_TEE_OBJECT) > + return "root"; > + > + if (!object->name) > + return "noname"; > + > + return object->name; > +} > + > +struct qcom_tee_object *allocate_qcom_tee_object(void); > +void free_qcom_tee_object(struct qcom_tee_object *object); > + > +/** > + * init_qcom_tee_object_user - Initialize an instance of qcom_tee_object. > + * @object: object being initialized. > + * @ot: type of object. > + * @ops: sets of callback opeartions. > + * @fmt: object name. > + */ > +int init_qcom_tee_object_user(struct qcom_tee_object *object, enum qcom_tee_object_type ot, > + struct qcom_tee_object_operations *ops, const char *fmt, ...); > + > +int get_qcom_tee_object(struct qcom_tee_object *object); > +void put_qcom_tee_object(struct qcom_tee_object *object); > + > +#endif /* __QCOM_OBJECT_INVOKE_H */ > > -- > 2.34.1 > -- With best wishes Dmitry

1 year, 2 months

Re: [PATCH RFC 0/3] Implement Qualcomm TEE IPC and ioctl calls

by Dmitry Baryshkov

On Tue, Jul 02, 2024 at 10:57:35PM GMT, Amirreza Zarrabi wrote: > Qualcomm TEE hosts Trusted Applications (TAs) and services that run in > the secure world. Access to these resources is provided using MinkIPC. > MinkIPC is a capability-based synchronous message passing facility. It > allows code executing in one domain to invoke objects running in other > domains. When a process holds a reference to an object that lives in > another domain, that object reference is a capability. Capabilities > allow us to separate implementation of policies from implementation of > the transport. > > As part of the upstreaming of the object invoke driver (called SMC-Invoke > driver), we need to provide a reasonable kernel API and UAPI. The clear > option is to use TEE subsystem and write a back-end driver, however the > TEE subsystem doesn't fit with the design of Qualcomm TEE. > > Does TEE subsystem fit requirements of a capability based system? > ----------------------------------------------------------------- > In TEE subsystem, to invoke a function: > - client should open a device file "/dev/teeX", > - create a session with a TA, and > - invoke the functions in that session. > > 1. The privilege to invoke a function is determined by a session. If a > client has a session, it cannot share it with other clients. Even if > it does, it is not fine-grained enough, i.e. either all accessible > functions/resources in a session or none. Assume a scenario when a client > wants to grant a permission to invoke just a function that it has the rights, > to another client. > > The "all or nothing" for sharing sessions is not in line with our > capability system: "if you own a capability, you should be able to grant > or share it". Can you please be more specific here? What kind of sharing is expected on the user side of it? > 2. In TEE subsystem, resources are managed in a context. Every time a > client opens "/dev/teeX", a new context is created to keep track of > the allocated resources, including opened sessions and remote objects. Any > effort for sharing resources between two independent clients requires > involvement of context manager, i.e. the back-end driver. This requires > implementing some form of policy in the back-end driver. What kind of resource sharing? > 3. The TEE subsystem supports two type of memory sharing: > - per-device memory pools, and > - user defined memory references. > User defined memory references are private to the application and cannot > be shared. Memory allocated from per-device "shared" pools are accessible > using a file descriptor. It can be mapped by any process if it has > access to it. This means, we cannot provide the resource isolation > between two clients. Assume a scenario when a client wants to allocate a > memory (which is shared with TEE) from an "isolated" pool and share it > with another client, without the right to access the contents of memory. This doesn't explain, why would it want to share such memory with another client. > 4. The kernel API provided by TEE subsystem does not support a kernel > supplicant. Adding support requires an execution context (e.g. a > kernel thread) due to the TEE subsystem design. tee_driver_ops supports > only "send" and "receive" callbacks and to deliver a request, someone > should wait on "receive". There is nothing wrong here, but maybe I'm misunderstanding something. > We need a callback to "dispatch" or "handle" a request in the context of > the client thread. It should redirect a request to a kernel service or > a user supplicant. In TEE subsystem such requirement should be implemented > in TEE back-end driver, independent from the TEE subsystem. > > 5. The UAPI provided by TEE subsystem is similar to the GPTEE Client > interface. This interface is not suitable for a capability system. > For instance, there is no session in a capability system which means > either its should not be used, or we should overload its definition. General comment: maybe adding more detailed explanation of how the capabilities are aquired and how they can be used might make sense. BTW. It might be my imperfect English, but each time I see the word 'capability' I'm thinking that some is capable of doing something. I find it hard to use 'capability' for the reference to another object. > > Can we use TEE subsystem? > ------------------------- > There are workarounds for some of the issues above. The question is if we > should define our own UAPI or try to use a hack-y way of fitting into > the TEE subsystem. I am using word hack-y, as most of the workaround > involves: > > - "diverging from the definition". For instance, ignoring the session > open and close ioctl calls or use file descriptors for all remote > resources (as, fd is the closet to capability) which undermines the > isolation provided by the contexts, > > - "overloading the variables". For instance, passing object ID as file > descriptors in a place of session ID, or > > - "bypass TEE subsystem". For instance, extensively rely on meta > parameters or push everything (e.g. kernel services) to the back-end > driver, which means leaving almost all TEE subsystem unused. > > We cannot take the full benefits of TEE subsystem and may need to > implement most of the requirements in the back-end driver. Also, as > discussed above, the UAPI is not suitable for capability-based use cases. > We proposed a new set of ioctl calls for SMC-Invoke driver. > > In this series we posted three patches. We implemented a transport > driver that provides qcom_tee_object. Any object on secure side is > represented with an instance of qcom_tee_object and any struct exposed > to TEE should embed an instance of qcom_tee_object. Any, support for new > services, e.g. memory object, RPMB, userspace clients or supplicants are > implemented independently from the driver. > > We have a simple memory object and a user driver that uses > qcom_tee_object. Could you please point out any user for the uAPI? I'd like to understand how does it from from the userspace point of view. > > Signed-off-by: Amirreza Zarrabi <quic_azarrabi(a)quicinc.com> > --- > Amirreza Zarrabi (3): > firmware: qcom: implement object invoke support > firmware: qcom: implement memory object support for TEE > firmware: qcom: implement ioctl for TEE object invocation > > drivers/firmware/qcom/Kconfig | 36 + > drivers/firmware/qcom/Makefile | 2 + > drivers/firmware/qcom/qcom_object_invoke/Makefile | 12 + > drivers/firmware/qcom/qcom_object_invoke/async.c | 142 +++ > drivers/firmware/qcom/qcom_object_invoke/core.c | 1139 ++++++++++++++++++ > drivers/firmware/qcom/qcom_object_invoke/core.h | 186 +++ > .../qcom/qcom_object_invoke/qcom_scm_invoke.c | 22 + > .../firmware/qcom/qcom_object_invoke/release_wq.c | 90 ++ > .../qcom/qcom_object_invoke/xts/mem_object.c | 406 +++++++ > .../qcom_object_invoke/xts/object_invoke_uapi.c | 1231 ++++++++++++++++++++ > include/linux/firmware/qcom/qcom_object_invoke.h | 233 ++++ > include/uapi/misc/qcom_tee.h | 117 ++ > 12 files changed, 3616 insertions(+) > --- > base-commit: 74564adfd3521d9e322cfc345fdc132df80f3c79 > change-id: 20240702-qcom-tee-object-and-ioctls-6f52fde03485 > > Best regards, > -- > Amirreza Zarrabi <quic_azarrabi(a)quicinc.com> > -- With best wishes Dmitry

1 year, 2 months

Re: [PATCH v5 2/9] scatterlist: Add a flag for the restricted memory

by Christian König

Am 27.06.24 um 05:21 schrieb Jason-JH Lin (林睿祥): > > On Wed, 2024-06-26 at 19:56 +0200, Daniel Vetter wrote: > > > > External email : Please do not click links or open attachments until > > you have verified the sender or the content. > > On Wed, Jun 26, 2024 at 12:49:02PM +0200, Christian König wrote: > > > Am 26.06.24 um 10:05 schrieb Jason-JH Lin (林睿祥): > > > > > > I think I have the same problem as the ECC_FLAG mention in: > > > > > > > > > > https://lore.kernel.org/linux-media/20240515-dma-buf-ecc-heap-v1-0-54cbbd04… > > > > > > > > I think it would be better to have the user configurable > > private > > > > > > information in dma-buf, so all the drivers who have the same > > > > > > requirement can get their private information from dma-buf > > directly > > > > > > and > > > > > > no need to change or add the interface. > > > > > > > > What's your opinion in this point? > > > > > > Well of hand I don't see the need for that. > > > > > > What happens if you get a non-secure buffer imported in your > > secure > > > > > device? > > > > > > > > We use the same mediatek-drm driver for secure and non-secure > > buffer. > > > > If non-secure buffer imported to mediatek-drm driver, it's go to > > the > > > > normal flow with normal hardware settings. > > > > > > > > We use different configurations to make hardware have different > > > > permission to access the buffer it should access. > > > > > > > > So if we can't get the information of "the buffer is allocated > > from > > > > restricted_mtk_cma" when importing the buffer into the driver, we > > won't > > > > be able to configure the hardware correctly. > > > > > > Why can't you get this information from userspace? > > > > Same reason amd and i915/xe also pass this around internally in the > > kernel, it's just that for those gpus the render and kms node are the > > same > > driver so this is easy. > > The reason I ask is that encryption here looks just like another parameter for the buffer, e.g. like format, stride, tilling etc.. So instead of this during buffer import: mtk_gem->secure = (!strncmp(attach->dmabuf->exp_name, "restricted", 10)); mtk_gem->dma_addr = sg_dma_address(sg->sgl); mtk_gem->size = attach->dmabuf->size; mtk_gem->sg = sg; You can trivially say during use hey this buffer is encrypted. At least that's my 10 mile high view, maybe I'm missing some extensive key exchange or something like that. > > > But on arm you have split designs everywhere and dma-buf > > import/export, so > > something else is needed. And neither current kms uapi nor > > protocols/extensions have provisions for this (afaik) because it > > works on > > the big gpus, and on android it's just hacked up with backchannels. > > > > So yeah essentially I think we probably need something like this, as > > much > > as it sucks. I see it somewhat similar to handling pcip2pdma > > limitations > > in the kernel too. > > > > Not sure where/how it should be handled though, and maybe I've missed > > something around protocols, in which case I guess we should add some > > secure buffer flags to the ADDFB2 ioctl. > > Thanks for your hint, I'll try to add the secure flag to the ADDFB2 > ioctl. If it works, I'll send the patch. Yeah, exactly what I would suggest as well. I'm not an expert for that part, but as far as I know we already have bunch of device specific tilling flags in there. Adding an MTK_ENCRYPTED flag should be trivial. Regards, Christian. > > Regards, > Jason-JH.Lin > > > -Sima > > ************* MEDIATEK Confidentiality Notice ******************** > The information contained in this e-mail message (including any > attachments) may be confidential, proprietary, privileged, or otherwise > exempt from disclosure under applicable laws. It is intended to be > conveyed only to the designated recipient(s). Any use, dissemination, > distribution, printing, retaining or copying of this e-mail (including its > attachments) by unintended recipient(s) is strictly prohibited and may > be unlawful. If you are not an intended recipient of this e-mail, or believe > that you have received this e-mail in error, please notify the sender > immediately (by replying to this e-mail), delete any and all copies of > this e-mail (including any attachments) from your system, and do not > disclose the content of this e-mail to any other person. Thank you!

1 year, 2 months

dma_buf_detach lockdep splat

by Thomas Hellström

Hi! I'm seeing the below lockdep splat 1) with the xe driver in an imported dma-buf object destruction path. It's not because we hold the dma_resv lock at that point, but rather because we hold *another* dma_resv lock at that point, and the dma_resv detach happens when the object is idle, in this case it was idle at the final put(), and dma_buf_detach() is called in the putting process. Holding another dma-buf lock might happen as part of drm_exec_unlock_all, or simply if the wider vm dma_resv was held at object put time, so it's not an uncommon pattern, even if the drm_exec instance can be fixed by putting all bos after unlocking them all. Two solutions coming to mind here: 1) Provide a dma_buf_detach_locked() 2) Have TTM always take the delayed delete path for imported dma-buf objects. I'd prefer 1) since I think the correct place to call this is in the TTM callback delete_mem_notify() where the bo is already locked, and I figure non-TTM gem backends may come to suffer from the same problem. Opinions, suggestions? [1] [ 99.136161] ============================================ [ 99.136162] WARNING: possible recursive locking detected [ 99.136163] 6.10.0-rc2+ #6 Tainted: G U [ 99.136165] -------------------------------------------- [ 99.136166] glxgears:sh0/4675 is trying to acquire lock: [ 99.136167] ffff9967dcdd91a8 (reservation_ww_class_mutex){+.+.}- {3:3}, at: dma_buf_detach+0x3b/0xf0 [ 99.136184] but task is already holding lock: [ 99.136186] ffff9967d8c145a8 (reservation_ww_class_mutex){+.+.}- {3:3}, at: drm_exec_lock_obj+0x49/0x2b0 [drm_exec] [ 99.136191] other info that might help us debug this: [ 99.136192] Possible unsafe locking scenario: [ 99.136194] CPU0 [ 99.136194] ---- [ 99.136195] lock(reservation_ww_class_mutex); [ 99.136197] lock(reservation_ww_class_mutex); [ 99.136199] *** DEADLOCK *** [ 99.136199] May be due to missing lock nesting notation [ 99.136200] 5 locks held by glxgears:sh0/4675: [ 99.136202] #0: ffff9967d8c104c8 (&xef->vm.lock){+.+.}-{3:3}, at: xe_file_close+0xde/0x1c0 [xe] [ 99.136272] #1: ffff9967d5bb7480 (&vm->lock){++++}-{3:3}, at: xe_vm_close_and_put+0x161/0x9b0 [xe] [ 99.136350] #2: ffff9967ef88a970 (&val->lock){.+.+}-{3:3}, at: xe_validation_ctx_init+0x6d/0x70 [xe] [ 99.136440] #3: ffffbd6a085577b8 (reservation_ww_class_acquire){+.+.}-{0:0}, at: xe_vma_destroy_unlocked+0x7f/0xe0 [xe] [ 99.136546] #4: ffff9967d8c145a8 (reservation_ww_class_mutex){+.+.}-{3:3}, at: drm_exec_lock_obj+0x49/0x2b0 [drm_exec] [ 99.136552] stack backtrace: [ 99.136553] CPU: 10 PID: 4675 Comm: glxgears:sh0 Tainted: G U 6.10.0-rc2+ #6 [ 99.136555] Hardware name: ASUS System Product Name/PRIME B560M-A AC, BIOS 2001 02/01/2023 [ 99.136557] Call Trace: [ 99.136558] <TASK> [ 99.136560] dump_stack_lvl+0x77/0xb0 [ 99.136564] __lock_acquire+0x1232/0x2160 [ 99.136569] lock_acquire+0xcb/0x2d0 [ 99.136570] ? dma_buf_detach+0x3b/0xf0 [ 99.136574] ? __lock_acquire+0x417/0x2160 [ 99.136577] __ww_mutex_lock.constprop.0+0xd0/0x13b0 [ 99.136580] ? dma_buf_detach+0x3b/0xf0 [ 99.136584] ? dma_buf_detach+0x3b/0xf0 [ 99.136588] ? ww_mutex_lock+0x2b/0x90 [ 99.136590] ww_mutex_lock+0x2b/0x90 [ 99.136592] dma_buf_detach+0x3b/0xf0 [ 99.136595] drm_prime_gem_destroy+0x2f/0x40 [drm] [ 99.136638] xe_ttm_bo_destroy+0x32/0x220 [xe] [ 99.136734] ? __mutex_unlock_slowpath+0x3a/0x290 [ 99.136738] drm_exec_unlock_all+0xa1/0xd0 [drm_exec] [ 99.136741] drm_exec_fini+0x12/0xb0 [drm_exec] [ 99.136743] xe_validation_ctx_fini+0x15/0x40 [xe] [ 99.136848] xe_vma_destroy_unlocked+0xb1/0xe0 [xe] [ 99.136954] xe_vm_close_and_put+0x41a/0x9b0 [xe] [ 99.137056] ? xa_find+0xe3/0x1e0 [ 99.137060] xe_file_close+0x10a/0x1c0 [xe] [ 99.137157] drm_file_free+0x22a/0x280 [drm] [ 99.137193] drm_release_noglobal+0x22/0x70 [drm] [ 99.137227] __fput+0xf1/0x2d0 [ 99.137231] task_work_run+0x59/0x90 [ 99.137235] do_exit+0x330/0xb40 [ 99.137238] do_group_exit+0x36/0xa0 [ 99.137241] get_signal+0xbd2/0xbe0 [ 99.137245] arch_do_signal_or_restart+0x3e/0x240 [ 99.137249] syscall_exit_to_user_mode+0x1e7/0x290 [ 99.137252] do_syscall_64+0xa1/0x180 [ 99.137255] ? _raw_spin_unlock+0x23/0x40 [ 99.137257] ? look_up_lock_class+0x6f/0x120 [ 99.137261] ? __lock_acquire+0x417/0x2160 [ 99.137264] ? lock_acquire+0xcb/0x2d0 [ 99.137266] ? __set_task_comm+0x28/0x1e0 [ 99.137268] ? find_held_lock+0x2b/0x80 [ 99.137271] ? __set_task_comm+0xe1/0x1e0 [ 99.137273] ? lock_release+0xca/0x290 [ 99.137277] ? __do_sys_prctl+0x245/0xab0 [ 99.137279] ? lockdep_hardirqs_on_prepare+0xde/0x190 [ 99.137281] ? syscall_exit_to_user_mode+0xb0/0x290 [ 99.137284] ? do_syscall_64+0xa1/0x180 [ 99.137286] ? cpuset_cpus_allowed+0x36/0x140 [ 99.137289] ? find_held_lock+0x2b/0x80 [ 99.137291] ? find_held_lock+0x2b/0x80 [ 99.137294] ? __sched_setaffinity+0x78/0x240 [ 99.137297] ? kfree+0xe2/0x310 [ 99.137301] ? kfree+0x202/0x310 [ 99.137303] ? __sched_setaffinity+0x78/0x240 [ 99.137305] ? __x64_sys_sched_setaffinity+0x69/0xb0 [ 99.137307] ? kfree+0xe2/0x310 [ 99.137310] ? lockdep_hardirqs_on_prepare+0xde/0x190 [ 99.137312] ? syscall_exit_to_user_mode+0xb0/0x290 [ 99.137315] ? do_syscall_64+0xa1/0x180 [ 99.137317] ? trace_hardirqs_off+0x4b/0xc0 [ 99.137321] ? clear_bhb_loop+0x45/0xa0 [ 99.137325] ? clear_bhb_loop+0x45/0xa0 [ 99.137327] ? clear_bhb_loop+0x45/0xa0 [ 99.137330] entry_SYSCALL_64_after_hwframe+0x76/0x7e [ 99.137333] RIP: 0033:0x7fda70ee6169 [ 99.137351] Code: Unable to access opcode bytes at 0x7fda70ee613f. [ 99.137352] RSP: 002b:00007fda5fdffc80 EFLAGS: 00000246 ORIG_RAX: 00000000000000ca [ 99.137354] RAX: fffffffffffffe00 RBX: 0000000000000000 RCX: 00007fda70ee6169 [ 99.137356] RDX: 0000000000000000 RSI: 0000000000000189 RDI: 0000564a96f45b30 [ 99.137358] RBP: 00007fda5fdffcb0 R08: 0000000000000000 R09: 00000000ffffffff [ 99.137359] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000 [ 99.137360] R13: 0000000000000000 R14: 0000000000000000 R15: 0000564a96f45b30 [ 99.137365] </TASK>

1 year, 2 months

[PATCH v12 0/7] iio: new DMABUF based API v12

by Paul Cercueil

Hi Jonathan, Here's the v12 of my patchset that introduces DMABUF support to IIO. Apart from a small documentation fix, it reverts to using mutex_lock/mutex_unlock in one particular place, which used cleanup GOTOs (which don't play well with scope-managed cleanups). Changelog: - [3/7]: - Revert to mutex_lock/mutex_unlock in iio_buffer_attach_dmabuf(), as it uses cleanup GOTOs - [6/7]: - "obtained using..." -> "which can be obtained using..." This is based on next-20240619. Cheers, -Paul Paul Cercueil (7): dmaengine: Add API function dmaengine_prep_peripheral_dma_vec() dmaengine: dma-axi-dmac: Implement device_prep_peripheral_dma_vec iio: core: Add new DMABUF interface infrastructure iio: buffer-dma: Enable support for DMABUFs iio: buffer-dmaengine: Support new DMABUF based userspace API Documentation: iio: Document high-speed DMABUF based API Documentation: dmaengine: Document new dma_vec API Documentation/driver-api/dmaengine/client.rst | 9 + .../driver-api/dmaengine/provider.rst | 10 + Documentation/iio/iio_dmabuf_api.rst | 54 +++ Documentation/iio/index.rst | 1 + drivers/dma/dma-axi-dmac.c | 40 ++ drivers/iio/Kconfig | 1 + drivers/iio/buffer/industrialio-buffer-dma.c | 178 ++++++- .../buffer/industrialio-buffer-dmaengine.c | 62 ++- drivers/iio/industrialio-buffer.c | 459 ++++++++++++++++++ include/linux/dmaengine.h | 33 ++ include/linux/iio/buffer-dma.h | 31 ++ include/linux/iio/buffer_impl.h | 33 ++ include/uapi/linux/iio/buffer.h | 22 + 13 files changed, 913 insertions(+), 20 deletions(-) create mode 100644 Documentation/iio/iio_dmabuf_api.rst -- 2.43.0

1 year, 2 months

Re: [PATCH v5 2/9] scatterlist: Add a flag for the restricted memory

by Christian König

Am 27.06.24 um 05:17 schrieb Jason-JH Lin (林睿祥): > > On Wed, 2024-06-26 at 12:49 +0200, Christian König wrote: > > > > External email : Please do not click links or open attachments until > > you have verified the sender or the content. > > Am 26.06.24 um 10:05 schrieb Jason-JH Lin (林睿祥): > > > > > > > > > In the step 3), we need to verify the dma-buf is allocated from > > > > > "restricted_mtk_cma", but there is no way to pass the > > > secure flag > > > > > or > > > > > private data from userspace to the import interface in DRM > > > driver. > > > > > > > > Why do you need to verify that? > > > > > > I need to know the imported buffer is allocated from restricted cma > > > and > > > mark it as a secure buffer in mediatek-drm driver. Then, I will add > > > some configuration to the hardware if the buffer is secure buffer, > > > so > > > that it can get the permission to access the secure buffer. > > > > Yeah so far that makes sense. This is basically what other drivers do > > with secure buffers as well. > > > > But why do you want the kernel to transport that information? Usually > > drivers get the information from userspace what to do with a buffer. > > > > In other words the format, stride, tilling and also if it's a secure > > buffer or not comes from userspace. > > > > Thanks for your clear explanation. > I think this is what I want, but I can't find any DRM interface to pass > the secure flag from userspace. Well stuff like that is usually device driver specific. So you should probably use something device specific which tells the driver that this buffer is encrypted. > > What the hardware usually handles internally is things like > > encryption keys, but you eventually get the information where to look > > for the key from userspace as well. > > > > Handling inside the kernel would only be necessary if userspace could > > for example crash the system with invalid parameters. But for > > encryption that is usually not the case. > > > > Yes, that's true. > > > > > > > > > > So I can only verify it like this now: > > > > > struct drm_gem_object *mtk_gem_prime_import_sg_table(struct > > > > > drm_device > > > > > *dev, struct dma_buf_attachment *attach, struct sg_table *sg) > > > > > { > > > > > struct mtk_gem_obj *mtk_gem; > > > > > > > > > > /* check if the entries in the sg_table are contiguous */ > > > > > if (drm_prime_get_contiguous_size(sg) < > > > attach->dmabuf->size) { > > > > > DRM_ERROR("sg_table is not contiguous"); > > > > > return ERR_PTR(-EINVAL); > > > > > } > > > > > mtk_gem = mtk_gem_init(dev, attach->dmabuf->size); > > > > > if (IS_ERR(mtk_gem)) > > > > > return ERR_CAST(mtk_gem); > > > > > > > > > > + mtk_gem->secure = (!strncmp(attach->dmabuf->exp_name, > > > > > "restricted", > > > > > 10)); > > > > > mtk_gem->dma_addr = sg_dma_address(sg->sgl); > > > > > mtk_gem->size = attach->dmabuf->size; > > > > > mtk_gem->sg = sg; > > > > > > > > > > return &mtk_gem->base; > > > > > } > > > > > > > > Complete NAK from my side to that approach. Importing of a DMA- > > > buf > > > > should be independent of the exporter. > > > > > > > > What you could do is to provide the secure buffer from a device > > > and > > > > not a device heap. > > > > > > > > > > You mean I should allocate buffer in mediate-drm driver not > > > userspace? > > > > Well that depends. The question is if you have multiple drivers which > > needs to work with this secure buffer? > > > > If yes then you should have a general allocation heap for it. If no > > then the buffers could as well be allocated from the driver interface > > directly. > > > > Yes, this buffer needs work with GPU and DRM drivers, so this general > "restricted_mtk_cma" will allocated in userspace, then being passed to > GPU and DRM. Well do you really need a separate heap for that? In other words is only a certain part of the system memory capable of being encrypted? Or would the "normal" CMA heap do as well and you only need to setup your hardware properly for encryption? Additional to that in most other drivers buffer sharing and encryption are two separate things. In other words other drivers do something like this: 1. Allocate the buffer. 2. Import the buffer using DRM_IOCTL_PRIME_FD_TO_HANDLE. 3. Set additional buffer properties, e.g. format, stride, tilling, if it's secure, which encryption key to use, where to map etc... So as far as I can see the problem you are facing is that you try to mangle everything into DRM_IOCTL_PRIME_FD_TO_HANDLE. Why not make that a separate IOCTL? I mean we intentionally don't provide things like format, stride, tilling etc.. to DRM_IOCTL_PRIME_FD_TO_HANDLE. Encryption is just another of those parameters. Regards, Christian. > > > > I just have modified this to userspace by the comment here: > > > > > > > https://patchwork.kernel.org/project/linux-mediatek/patch/20240403102701.36… > > > > > > > > I think I have the same problem as the ECC_FLAG mention in: > > > > > > > > > > > > > > https://lore.kernel.org/linux-media/20240515-dma-buf-ecc-heap-v1-0-54cbbd04… > > > > > > > > > > I think it would be better to have the user configurable > > > private > > > > > information in dma-buf, so all the drivers who have the same > > > > > requirement can get their private information from dma-buf > > > directly > > > > > and > > > > > no need to change or add the interface. > > > > > > > > > > What's your opinion in this point? > > > > > > > > Well of hand I don't see the need for that. > > > > > > > > What happens if you get a non-secure buffer imported in your > > > secure > > > > device? > > > > > > We use the same mediatek-drm driver for secure and non-secure > > > buffer. > > > If non-secure buffer imported to mediatek-drm driver, it's go to > > > the > > > normal flow with normal hardware settings. > > > > > > We use different configurations to make hardware have different > > > permission to access the buffer it should access. > > > > > > So if we can't get the information of "the buffer is allocated from > > > restricted_mtk_cma" when importing the buffer into the driver, we > > > won't > > > be able to configure the hardware correctly. > > > > Why can't you get this information from userspace? > > As I mentioned here: > > https://patchwork.kernel.org/project/linux-mediatek/cover/20240525232928.55… > > I tried some DRM interfaces using buffer FD and arg->flag as > parameters, but it didn't work. So I ask for your help here. > > But I think I should find DRM maintainer to add the secure flag to DRM > interface now. > > Regards, > Jason-JH.Lin > > ************* MEDIATEK Confidentiality Notice ******************** > The information contained in this e-mail message (including any > attachments) may be confidential, proprietary, privileged, or otherwise > exempt from disclosure under applicable laws. It is intended to be > conveyed only to the designated recipient(s). Any use, dissemination, > distribution, printing, retaining or copying of this e-mail (including its > attachments) by unintended recipient(s) is strictly prohibited and may > be unlawful. If you are not an intended recipient of this e-mail, or believe > that you have received this e-mail in error, please notify the sender > immediately (by replying to this e-mail), delete any and all copies of > this e-mail (including any attachments) from your system, and do not > disclose the content of this e-mail to any other person. Thank you!

1 year, 2 months

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

Linaro-mm-sig