users August 2015

users@lists.genode.org

15 participants
14 discussions

missing pci_drv driver
by Guido Witmond 09 Aug '15

09 Aug '15

Hello Genodians, I'm about to test your new tg3 driver on my system. However, I cannot get genode to compile. I've got the latest genodelabs/genode repos checked out. This is the part that's breaking: diff --git a/repos/ports/run/lighttpd.run b/repos/ports/run/lighttpd.run index 60b9937..8148fe9 100644 --- a/repos/ports/run/lighttpd.run +++ b/repos/ports/run/lighttpd.run @@ -6,6 +6,7 @@ set build_components { core init + drivers/pci drivers/nic drivers/timer app/lighttpd @@ -41,7 +42,17 @@ append config { <provides><service name="Timer"/></provides> </start>} -append_platform_drv_config +#append_platform_drv_config +append_if [have_spec pci] config { + <start name="pci_drv"> + <resource name="RAM" quantum="4M"/> + <provides><service name="PCI"/></provides> + <config> + <policy label="nic_drv"> + <pci bus="0x10" device="0" function="0"/> + </policy> + </config> + </start>} append config { <start name="nic_drv"> In short: I need the pci driver to attach the nic_drv to the correct device, one of three available ethernet cards. It gives this error: make run/lighttpd ... Error: target 'drivers/pci' does not exist ... Where do I get the pci-driver from, it's missing in repos/os/src/drivers Cheers, Guido.

2 3

TrustZone: forcing instant switch from secure to normal world?
by David Goltzsche 04 Aug '15

04 Aug '15

Hi Genode team & list, in my master's thesis I built a framework, which enables cross-world-communication between the normal and secure world. This was based on your tz_vmm demo. If this sounds familiar: I already posted a few questions on this list related to my thesis ;) My framework relied on periodic world switches triggered by Genode's scheduler (The normal world is running within a Genode thread, so the scheduler prevents both worlds from starving). Of course, simply waiting for world switches leads to high latencies and a low throughput. To speed things up, we already improved the communication from normal to secure world: The normal world instantly hands over control to the secure world utilizing the smc instruction, when a message is available. We extended the hypercall handler in the Vmm class to notify the receiver via a semaphore. Unfortunately, forcing a world switch from secure to normal world does not seem to be that easy: Genode's monitor mode implementation cannot handle entries from the secure world. Hence, usage of the smc instruction is not possible. So, my approach would be using software interrupts which sould be routed to the normal world. I played around with the svc (supervisor call) instruction, but I don't know how to route this to the secure world. Any hints? Thanks in advance David -- M.Sc. David Goltzsche Institute of Operating Systems and Computer Networks Distributed Systems Group TU Braunschweig www: https://www.ibr.cs.tu-bs.de/users/goltzsch mail: goltzsche@...290...

3 7

tg3 driver works.
by Guido Witmond 04 Aug '15

04 Aug '15

Dear Genodians, Last weekend I compiled Foc_x86_64 and tested it with the onboard ethernet cards of my server box. It worked! It worked just as good as the Intel e1000 driver. (Meaning, after a short while, all throughput ceased. :-( But that is eiter Foc or Lightttpd) The good news: this makes Genode a viable platform for many server platform that use the Marvel chipset. Thanks! Cheers, Guido.

2 1

Data Communication btw Normal and Secure World in IMX53_QSB
by JaeminPark 03 Aug '15

03 Aug '15

I’m currently developing a simple trustzone device driver running in Normal World to communicate with Genode in Secure World. I’m using IMX53 QSB, Genode 15.02 as Secure OS, and customized Linux as OS for Normal World. In Normal World, my simple device driver calls smc instruction inside module_init(). The instruction is invoked with the command structure (named by “struct smc_cmd”). The physical address of "struct smc_cmd" is passing to Genode via _vm->state()->r1. There exists two physical addresses inside “struct smc_cmd": one for request buffer and the other for response buffer where the response buffer is allocated by kzalloc(). For the address conversion, I used virt_to_phys(). In Secure World (Genode), I’m modifying the tz_vmm source code for this experiment. I’m succeed in reading the request buffer from Normal World by directly accessing the physical address of the request buffer : PINF("req_buf = %s (0x%x)", (char*)((struct smc_cmd*)cmd_addr)->req_buf_phys, ((struct smc_cmd*)cmd_addr)->req_buf_phys); Ridiculously and Unfortunately, I’m in trouble to write some characters in response buffer (whose address is the physical address of Normal World). I used Genode::memcpy() to write the characters as follows where cmd_addr is the physical address of the command structure passed from Normal World : /* acquire data to be written into memory space of normal world */ Genode::memcpy(tmp, (const char*) "From Secure World", strlen("From Secure World")); tmp[17] = '\0'; /* write a data into resp_buf_phys */ Genode::memcpy((char*)((struct smc_cmd*)cmd_addr)->resp_buf_phys, (const char*) tmp, sizeof(tmp)); /* printout contents of memory for debug */ PINF("res_buf = %s (0x%x)", (char*)((struct smc_cmd*)cmd_addr)->resp_buf_phys, ((struct smc_cmd*)cmd_addr)->resp_buf_phys); It seems that the writing operation in Secure World is successful since the last PINF() shows me the “From Secure World” in console. However, when I read the address (converted from the physical address using phys_to_virt()) in my device driver, I can only see the empty string (initial status). My questions are : Do I need another approach to write the physical memory of Normal World in the side of Genode? Is there any simple example or any volunteer to share an example for this kind of operation? Is there any comment to implement this kind of example?

3 2

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

users August 2015