Thanks a lot, Jiuyue, This is really insightful. Let me see how it goes.
On Fri, Jun 13, 2014 at 4:41 AM, 马久跃 via gem5-users <[email protected]> wrote: > Sorry for the mistake in previous reply. > > The reason of invalid address error in 2.6.28.4 kernel was CPUID > LongModeAddressSize(eax:0x80000008) not implement in gem5. > > *linux-2.8.28.4/arch/x86/mm/ioremap.c* > > 27 static inline int phys_addr_valid(unsigned long addr) > 28 { > 29 return addr < (1UL << boot_cpu_data.x86_phys_bits); > 30 } > ... > 191 static void __iomem *__ioremap_caller(resource_size_t phys_addr, > 192 unsigned long size, unsigned long prot_val, void > *caller) > ... > 209 if (!phys_addr_valid(phys_addr)) { * <- > ioremap > invalid address error here* > 210 printk(KERN_WARNING "ioremap: invalid physical address > %llx\n", > 211 (unsigned long long)phys_addr); > 212 WARN_ON_ONCE(1); > 213 return NULL; > 214 } > > > > *linux-2.6.28.4/arch/x86/kernel/cpu/common.c* > > 475 #ifdef CONFIG_X86_64 > 476 if (c->extended_cpuid_level >= 0x80000008) { > 477 u32 eax = cpuid_eax(0x80000008); > 478 > 479 c->x86_virt_bits = (eax >> 8) & 0xff; > 480 c->x86_phys_bits = eax & 0xff; > * <- x86_phys_bits** initialized **here, using **CPUID > LongModeAddressSize* > 481 } > 482 #endif > > > The attached two patch should fix this problem, one for > LongModeAddressSize, and the other for x86-ethernet config. > cpuid patch return 0x0000ffff for LongModeAddressSize, I'm not sure > whether it was right, but it works for me. > > > > Jiuyue Ma > > > ------------------------------ > To: [email protected] > Date: Fri, 13 Jun 2014 08:33:29 +0800 > > Subject: Re: [gem5-users] how to enable Ethernet device for x86 full system > From: [email protected] > > Try to connect "x86_sys.membus.default" port to "x86_sys.bridge.slave" > instead of badaddr_responder (configs/common/FSConfig.py:53-55,334-340), > this should solve the invalid address error. > You may also need change ethernet device's default InterruptLine (0x1e) to > a smaller value (e.g. less than 0x10), otherwise X86 I/O APIC may not > work properly. > > Jiuyue Ma > > > ------------------------------ > Date: Thu, 12 Jun 2014 08:00:34 -0700 > To: [email protected]; [email protected] > Subject: Re: [gem5-users] how to enable Ethernet device for x86 full system > From: [email protected] > > I encountered the same issue. I think this should be a supported feature. > > -Pete > > > > On Thu, Jun 12, 2014 at 6:55 AM, Veydan Wu via gem5-users < > [email protected]> wrote: > > HI, All, > > I want to enable an Ethernet device to x86 full system simulation. After > searching the mail archive, I followed the ARM way, added the device in > dev/x86/Pc.py just like Arm Realview, and added code to attachIO(). The > config.ini file showed the device is there and connected on iobus. > > But when booting the kernel (2.6.28.4) using atomic CPU, there was a > ioremap invalid address error. And after boot, the system can't see the > device (ifconfig eth0 shows nothing). > > Has anyone successfully made this work? I am not familiar with the x86 IO > configuration. Any hints would be really helpful. Thanks a lot. > > -- > Regards, > > Veydan > > _______________________________________________ > gem5-users mailing list > [email protected] > http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users > > > > _______________________________________________ gem5-users mailing list > [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users > > _______________________________________________ gem5-users mailing list > [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users > > _______________________________________________ > gem5-users mailing list > [email protected] > http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users > -- Regards, Veydan
_______________________________________________ gem5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
