users

users@lists.genode.org

2285 discussions

Assigning PCI device to device_pd
by Alexander Senier 22 Nov '16

22 Nov '16

Hi all, I'm building a toy driver to experiment with the details of Intel integrated graphics. My experiments are based on Genode 16.08 (master) using NOVA. Getting a capability to the IGD device (0:2:0) from the platform driver works, so does reading/writing PCI config space, mapping BAR0 and accessing MMIO registers. Next, I want to test DMA from the IGD device by 1. Allocating DMA memory (via alloc_dma_buffer()) and attaching it to the driver address space 2. Mapping the aperture (BAR2) 3. Pointing the GTT to the physical address of the DMA memory allocated in 1. 4. Writing some data to the DMA memory 5. Reading that data back through the aperture Unfortunately, I'm stuck with that test. Apparently the IOMMU mappings are not set up correctly and NOVA emits respective IOMMU faults: [ 0] DMAR:0xffffffff81034040 FRR:0 FR:0x5 BDF:0:2:0 FI:0x041d4000 The BDF is the expected IGD device and the fault address matches the physical address of the allocated DMA memory (and the PTE set up in the GTT). When instrumenting device_pd, I see that attach_dma_mem() is called for the correct physical address and size. However, assign_device() never gets called for 0:2:0. The only place I see where assign_device() is used in the code, is in config_write() when writing PCI_CMD_DMA to PCI_CMD_REG. However, when doing such a config_write() manually in my component, I get an Out_of_metadate exception. How can I assign the PCI device to my drivers device_pd? See [1] for example code. Thanks a lot! Cheers, Alex [1] https://github.com/senier/gpu_kernel/tree/master/src/app/hello_gpu

2 2

Re: RPC interface to get random numbers
by Duss Pirmin 17 Nov '16

17 Nov '16

Spam detection software, running on the system "idefix.duss.local", has identified this incoming email as possible spam. The original message has been attached to this so you can view it or label similar future email. If you have any questions, see @@CONTACT_ADDRESS@@ for details. Content preview: Hello Norman Quoting Norman Feske <norman.feske@...1...>: > Hello Pirmin, > > thanks for posting your idea to the list. > > I wonder, have you considered to implement the random service as a > "Terminal" service that just produces a stream of random values? The > benefit would be that such a server can be combined with the existing > terminal infrastructure out of the box. For example, you could mount the > random server's session in the VFS (via a <terminal> node) without any > extra effort. [...] Content analysis details: (7.1 points, 5.0 required) pts rule name description ---- ---------------------- -------------------------------------------------- 0.2 CK_HELO_GENERIC Relay used name indicative of a Dynamic Pool or Generic rPTR 2.9 HELO_DYNAMIC_SPLIT_IP Relay HELO'd using suspicious hostname (Split IP) 3.6 RCVD_IN_PBL RBL: Received via a relay in Spamhaus PBL [92.106.180.179 listed in zen.spamhaus.org] 0.0 RCVD_IN_SORBS_DUL RBL: SORBS: sent directly from dynamic IP address [92.106.180.179 listed in dnsbl.sorbs.net] 0.4 RDNS_DYNAMIC Delivered to internal network by host with dynamic-looking rDNS

1 0

RPC interface to get random numbers
by Duss Pirmin 15 Nov '16

15 Nov '16

Spam detection software, running on the system "idefix.duss.local", has identified this incoming email as possible spam. The original message has been attached to this so you can view it or label similar future email. If you have any questions, see @@CONTACT_ADDRESS@@ for details. Content preview: Hello We have different components in our system that require random numbers. So our idea is to implement an RPC interface that allows one component to provide random numbers and an other to receive them. [...] Content analysis details: (7.1 points, 5.0 required) pts rule name description ---- ---------------------- -------------------------------------------------- 0.2 CK_HELO_GENERIC Relay used name indicative of a Dynamic Pool or Generic rPTR 2.9 HELO_DYNAMIC_SPLIT_IP Relay HELO'd using suspicious hostname (Split IP) 0.0 RCVD_IN_SORBS_DUL RBL: SORBS: sent directly from dynamic IP address [92.106.180.179 listed in dnsbl.sorbs.net] 3.6 RCVD_IN_PBL RBL: Received via a relay in Spamhaus PBL [92.106.180.179 listed in zen.spamhaus.org] 0.4 RDNS_DYNAMIC Delivered to internal network by host with dynamic-looking rDNS

2 1

memory reference monitor in SW
by Rijurekha Sen 11 Nov '16

11 Nov '16

I have an imx53 sabre tablet. I am trying to implement a reference monitor in SW, to create a trace for memory accesses from android running in NW. I do not need to monitor all memory accesses, but those corresponding to IPU and I2C. I am actually interested in the OV5642 camera module - and want to monitor in SW all memory read/writes corresponding to android camera activities running in NW. Going through the android device driver code, saw that IPU and I2C do the memory related operations and configurations for the camera. In http://www.mit.edu/afs.new/sipb.mit.edu/project/freebsd/head/sys/gnu/dts/ar…, which matches the IMX53 reference manual, the IPU and three I2C physical address ranges are mentioned. If I want to create a reference monitor for these physical memory addresses, should I mark these as secure memory? According to https://community.arm.com/thread/4852#15483, to just read/write NW pages, SW doesn’t need extra hardware configuration like TZASC, M4IF, TZPC or CSU. This is possibly what https://www.informatik.tu-darmstadt.de/fileadmin/user_upload/Group_TRUST/Pu… uses to check/modify android loaded in NW RAM. But my SW will need to know when NW is making a particular memory access, so the ability to passively read/write NW memory locations from SW isn’t useful. How can I mark multiple blocks of physical address ranges (corresponding to memory mapped IO and not SRAM/DRAM) as secure? The M4IF seems to split the DDR external memory into two blocks. Should I use TZASC, as according to genode documentation - "Due to experimentation, we were able to deduct the following insights. The TZASC controller on the platform is used to secure physical address ranges that are addressable through the Static Memory Controller (SMC). In principal, it should be possible to secure another memory controller by a TZASC too, but on the platform, it is restricted to the SMC. These physical address regions correspond to the I/O resources of peripheral devices, some SRAM, and flash memory.” Or do I need to configure the CSU? What is the way to configure TZASC or CSU, to make particular physical address ranges secure? When some physical memory is marked as secure, does NW access there cause a data abort? Does it automatically trap to monitor code, and SW can see what NW instruction caused it? I will need to record that access in SW, emulate that instruction in SW and give back control to NW. I understand that because IPU is needed for any display related activity (and accel/magnetometer/audio … all seem to be I2C devices), making those memory regions secure will cause android to fault even at booting. So I will possibly see the intended behavior, even without using the android camera. Thanks in advance for any suggestions on how to proceed. Thanks! Riju

1 1

booting android in normal world: imx53 sabre
by unknown＠example.com 10 Nov '16

10 Nov '16

Hi, I am using an IMX53 Sabre tablet, and can run the Genode demo. I need to reduce the TCB as much as possible, so ran the bare-bone kernel examples at https://github.com/finallyjustice/imx53qsb-code. for these, $make gave a kernel.imx which I wrote with dd to SD card. On power up, the board booted and ran the secure world code, and toggled with normal world, based on the program logic. There was no separate u-boot for these. I now need to boot android in NW, from the bare-bone kernel.imx in SW. I have the android images, the linux kernel uImage and u-boot.bin. Also, there is an android image in the e-mmc of the tablet, which the genode demo used. Can someone explain to me what I need to do - to boot android in NW from the bare-bone kernel in SW? (a) According to http://stackoverflow.com/questions/31589143/boot-linux-in-normal-world, https://community.nxp.com/thread/308152, https://community.nxp.com/thread/300696, I need to set some registers. I can see at https://github.com/finallyjustice/liboot-tz/blob/master/u-boot/arch/arm/lib…, how to do that. (b) Should I then write the kernel.imx of the compiled bare-bone program to SD card, followed by the android images, using dd with seek? The i.MX Android R10.4 User Guide partitioned the SD card into 5 partitions for data/boot/system/recovery etc. To boot android in NW, will the SD card need similar formatting and copying of files, and kernel.imx before that? Is there a way to not put the android image in the SD card, and instead load it from e-mmc? (c) http://stackoverflow.com/questions/29898139/load-two-kernel-images-using-u-… uses tftp to copy the image at the specified RAM locations, but how can I load the android image from the SD card/ emmc to a particular RAM location? I used ext2load mmc 0:1 0x70800000 image.bin for the genode demo. So here, will I need u-boot to load the kernel.imx and the android image to two separate addresses, start the kernel.imx with u-boot go, and then the "ldr pc, =android address" asm command followed by world switch will boot android? Thanks! Riju

1 0

Clang compatibilities
by Keiko Nakata 04 Nov '16

04 Nov '16

Hallo! Is the Genode framework source code compatible with Clang? Best regards, Keiko

2 2

Several sessions of the same type between server and client
by Igor 04 Nov '16

04 Nov '16

Hi list! I need to create several block sessions between ahci_drv (or part_blk) and my client. How it could be done? As far as I can understand, both aforementioned services can't work on a such manner. Did I miss something? Thanks in advance.

1 1

[tz_vmm demo]- execution time
by John David 03 Nov '16

03 Nov '16

Hello, I want to measure execution time of a function in the secure world of tz_vmm demo. I was wondering if I could use elapsed_ms( ) method in timer_session to measure t1 and t2 so that I can take the difference (t2 -t1). Is there any other way? Best regards, John

2 5

Supported ARM devices
by Yevgeny Lavrov 01 Nov '16

01 Nov '16

Hello. My current task is to bring Genode, preferably with seL4 on ARM. Can you please provide me a list of supported ARM devices, or recommend an ARM device that would be the best suited for this task. Thank you.

2 3

Unnecessary default-constructibility requirements for RPC types?
by Edgard Schmidt 26 Oct '16

26 Oct '16

Hi, as described in "Genode Foundations", all types which are used as argument or return types in RPC functions must have a default constructor. I do not think that such a limitation is really necessary. So I tried to write a proof of concept [1]. These modifications should allow RPC functions to pass most of the non-default-constructible objects which satisfy the rest of the requirements for RPC argument and return types. I extended the "hello_tutorial" with a little test code to show that it works [2]. Software patterns like RAII [3] are a possible reason for removing that limitation. Quite often classes cannot provide a useful default-constructor which does more than merely setting the object state to "invalid". Because often there are no useful default values. So if a class provides a default-constructor which just sets the object to "invalid", the code complexity increases since care has to be taken that the object is fed with data up to a specific point of time. Unfortunately, my implementation does not remove default-constructor calls completely: 1. Capabilities still need default-constructors. Msgbuf stores them directly in an array and I did not find the time to check changes of this storage behavior for possible side effects. 2. Output-only _arguments_ have to be default-constructible too. Please note, that non-default-constructible _return_ types are allowed. So this affects only RPC arguments which are explicitly declared as Output-only, for example by adding custom specializations of Rpc_direction. The reason for this limitation is that my Pod_tuple version still stores the RPC arguments directly. While input arguments are copy-constructed, it is hard to copy from an object that does not yet exist. For the second point, I see no simple solutions: we could create a simple byte buffer instead of a default-constructed output-only argument. Since "normal" RPC arguments have to be trivially copyable [4], a RPC function could easily write to this buffer, which would be accessible by the correctly type-casted reference or pointer. A minimalist version of boost::optional [5] or the like as member of Server_args could make sense. However, it would require more template specializations and case differentiations because the special RPC types Capability and Native_capability are not trivially copyable. I successfully tested the code with the repositories "hello_tutorial" and "demo" on base-linux. But if you decide to use my code contribution, please note that it is rather a proof of concept. There is room for improvements: * I tried to minimize the code modifications. So some refactoring may be useful: I expanded Meta::Empty by constructors and a typedef, I inserted possibly confusing forward declarations, and Meta::Pod_tuples and Meta::Pod_args could be renamed because they do not only accept POD types [6]. * I completely ignored the trace-related code. I do not know if there is a need of modifications. Thanks to Johannes Schlatow who gave me the idea of grappling with this topic. [1] https://github.com/3dik/genode/tree/nodef [2] https://github.com/3dik/genode/commit/c35a8d4a3166da74c46423f5e0b4a35933174… [3] https://en.wikipedia.org/wiki/Resource_Acquisition_Is_Initialization [4] http://en.cppreference.com/w/cpp/concept/TriviallyCopyable [5] http://www.boost.org/doc/libs/1_62_0/libs/optional/doc/html/boost_optional/… [6] http://en.cppreference.com/w/cpp/concept/PODType

2 2

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

users