SeaBIOS

seabios@seabios.org

4 participants
3801 discussions

[RFH] Using the memory error detector AddressSanitizer (ASan)
by Paul Menzel 26 Apr '15

26 Apr '15

Dear SeaBIOS folks, my goal is to build SeaBIOS with the “fast memory error detector” AddressSanitizer (ASan) [1]. As I did not find a way to override `CFLAGS` for building SeaBIOS, as they are defined with `:=` in `Makefile`, I added `-fsanitize=address` directly to `COMMONCFLAGS`. $ git diff diff --git a/Makefile b/Makefile index a4d945c..9a7897c 100644 --- a/Makefile +++ b/Makefile @@ -58,7 +58,7 @@ COMMONCFLAGS := -I$(OUT) -Isrc -Os -MD -g \ -m32 -march=i386 -mregparm=3 -mpreferred-stack-boundary=2 \ -minline-all-stringops -fomit-frame-pointer \ -freg-struct-return -ffreestanding -fno-delete-null-pointer-checks \ - -ffunction-sections -fdata-sections -fno-common -fno-merge-constants + -ffunction-sections -fdata-sections -fno-common -fno-merge-constants -fsanitize=address COMMONCFLAGS += $(call cc-option,$(CC),-nopie,) COMMONCFLAGS += $(call cc-option,$(CC),-fno-stack-protector,) COMMONCFLAGS += $(call cc-option,$(CC),-fno-stack-protector-all,) But, buidling SeaBIOS as a coreboot payload fails with a syntax error. […] Compiling (16bit) out/romlayout.o Building ld scripts Version: rel-1.8.0-24-g0b2165d-dirty-20150426_091353-myasrocke350m1 Fixed space: 0xe05b-0x10000 total: 8101 slack: 12 Percent slack: 0.1% 16bit size: 89760 32bit segmented size: 9184 32bit flat size: 65456 32bit flat init size: 126720 Lowmem size: 2176 f-segment var size: 1200 Linking out/rom16.o /usr/bin/ld.bfd.real:out/romlayout16.lds:8: syntax error Makefile:171: recipe for target 'out/rom16.o' failed make: *** [out/rom16.o] Error 1 The same happens when building SeaBIOS for QEMU/KVM. […] Building ld scripts Version: rel-1.8.0-24-g0b2165d-dirty-20150426_092029-myasrocke350m1 Fixed space: 0xe05b-0x10000 total: 8101 slack: 12 Percent slack: 0.1% 16bit size: 92640 32bit segmented size: 9728 32bit flat size: 82384 32bit flat init size: 200864 Lowmem size: 2176 f-segment var size: 1200 Linking out/rom16.o /usr/bin/ld.bfd.real:out/romlayout16.lds:9: syntax error Makefile:171: recipe for target 'out/rom16.o' failed make: *** [out/rom16.o] Error 1 Could you please suggest a way to build SeaBIOS with AddressSanitizer? Also I’m still curious how possible warnings/messages from AddressSanitzer will be shown to the user. Thanks, Paul [1] https://code.google.com/p/address-sanitizer/

1 0

Re: [SeaBIOS] [HELP] The program cause screen mess
by Kevin O'Connor 24 Apr '15

24 Apr '15

On Wed, Apr 22, 2015 at 02:20:51PM +0800, eJim Lee wrote: > Hi kevin: > If you want to repeat the bug, you can try: > 1. Go to "Control panel->Regional and Language > Options->Advanced->Language for non-Unicode programs" and select > "Chinese (PRC)". > 2. reboot > 3. run wbc > > I also tried the program on vmware and looked nice. But something > different with qemu about cursor: > On qemu with WinXP English: cursor is on screen but not on vmware. > > Hope these information can help you. > > Thanks! > eJim Lee > > 2015-04-21 23:41 GMT+08:00 eJim Lee <lyjforjob(a)gmail.com>: > > I have tried it without kvm and bug repeated: > > ./build/x86_64-softmmu/qemu-system-x86_64 -chardev stdio,id=seabios > > -device isa-debugcon,iobase=0x402,chardev=seabios -m 512 -hda > > tmp/xp.append -hdb tmp/data.append This issue doesn't seem like a vgabios issue. It doesn't appear that the program calls the vgabios and the program seems to function correctly when changing non-vgabios settings. So, it appears that something unknown is the root cause of the intermittent screen corruption. -Kevin

2 1

Re: [SeaBIOS] Problems booting legacy OSes on Chromebook Pixel 2015
by Kevin O'Connor 23 Apr '15

23 Apr '15

On Tue, Apr 21, 2015 at 07:26:27PM -0500, Charles Howell wrote: > (Apologies if this is the wrong list, but I'm not sure where else to post.) > > I'm attempting to boot FreeBSD 10.1 (amd64) on a new Chromebook Pixel (via > a USB CD) and it hangs as soon as the kernel is loaded into memory. > > I'd just assume it's a FreeBSD problem but I've also tried the latest > versions of OpenBSD and NetBSD with the same results. In all three cases, > the bootloader loads and I can interact with it, toggle options, read > machine memory layout, enable safe mode, etc. but as soon as the kernel is > loaded into memory and it tries to execute, the machine hangs hard and I > have to hold down power for a few seconds to shut it off. > > This leads me to believe it's a problem with SeaBIOS and/or its support for > this machine since I've booted BSDs on the Acer C720 with SeaBIOS just fine. > > I've also tried with a USB key with an installed OS already in case it was > related to reading the CD, but with the same results. The bootloader comes > up but no kernels will boot. > > These were all tested with SeaBIOS version -20150225_120401-build122-m2 as > shipped with the device. > > Can anyone offer any help? Unfortunately, it's very difficult to troubleshoot boot issues without access to the debug log. SeaBIOS currently only supports access to the debug log via a traditional serial port or via the "cbmem" in memory log. The cbmem log is only applicable if one can boot the machine to an OS that can run the cbmem tool. You could try taking the USB/CD image and running it on QEMU - just to verify that SeaBIOS doesn't have an issue with the image. (Though it's unlikely that is the cause.) Also, some of the more recent chromebooks have been shipping with SeaVGABIOS instead of the traditional Intel vgabios. This may confuse some kernels - you may want to verify that the machine has truly hung (as opposed to the video just hanging). Another option would be to try booting a Linux image - access to the cbmem console would be helpful even if booting Linux is not the goal. -Kevin

2 1

Extending USB device support
by Parker, John M 22 Apr '15

22 Apr '15

Hi I took some time and developed a serial console over USB (FTDI) for SeaBIOS. To make it work, I put in a few changes to the core USB module to allow my USB serial port driver to patch itself in to the USB stack. These changes implement a general mechanism for USB device driver extension that might have other applications, so I thought I would start a dialog on the subject. Here is a summary of the change (about 5 lines removed and 20 lines added): 1) defined a list structure in usb.h struct usb_driver_list_s { struct usb_driver_list_s *next; /* the list pointer */ int (*driver_detect)(struct usbdevice_s *usbdev, / struct usb_device_descriptor *dinfo); /* a function pointer */ } ; 2) Added a static list head pointer to usb.c static struct usb_driver_list_s *usb_driver_list; 3) Added an API function to populate the list with external usb drivers add_usb_device_handler(struct usb_driver_list_s *add_driver) 3) Modified configure_usb_device() to make it look for external drivers a) relocated the trap for an unhandled device b) call the handler init functions in the list until one indicates acceptance -----Original Message----- From: Kevin O'Connor [mailto:kevin@koconnor.net] Sent: Friday, April 03, 2015 10:58 AM To: Parker, John M Cc: seabios(a)seabios.org Subject: Re: [SeaBIOS] USB serial console support? On Fri, Apr 03, 2015 at 04:54:43PM +0000, Parker, John M wrote: > Hi > > This is a new-user question. > > I am working on a platform bring-up, specifically getting a port of > SeaBIOS sorted out. So far so good, the code has made it all the way > through to the USB enumeration. > > The current task is to bring up the boot menu on a 'serial over USB' > console. So: > Is a serial port boot menu supported? > Is there a serial over USB module in the works? Unfortunately, this feature isn't currently available. I'm not aware of anyone working on it. There are two parts to getting the above to work: forwarding keyboard and video over serial, and USB serial support. In the past, we've used sgabios ( https://code.google.com/p/sgabios/ ) to provide keyboard/video over traditional serial ports. However, sgabios wont work with USB serial ports (or even some PCI serial adapters) and it's unlikely that support will be added. Getting USB serial to work is a bit of a pain because there does not appear to be a standard for initializing USB serial adapters. One possibility would be to target only one device (and in particular, a device that something like a Beagle Bone Black could emulate would be very useful). So, to get the above to work, I think it would require both adding sgabios functionality into SeaBIOS and adding USB serial support into SeaBIOS. > > Thanks for your time > John Parker (john.m.parker(a)intel.com) -Kevin PS, please post in plain text only (no html) when sending to the mailing list.

1 0

Re: [SeaBIOS] Problems booting legacy OSes on Chromebook Pixel 2015
by Charles Howell 22 Apr '15

22 Apr '15

On Tue, Apr 21, 2015 at 7:26 PM, I wrote: > I'm attempting to boot FreeBSD 10.1 (amd64) on a new Chromebook Pixel (via > a USB CD) and it hangs as soon as the kernel is loaded into memory. > To eliminate the BSDs as the source of the problem, I just tried booting an ISO of Memtest86 (which boots v4.3 when UEFI is not present) and got the same results. The bootloader comes up, but as soon as it tries to actually load the program, the machine locks up ("Loading memtest...ready." and then lock up). If it makes any difference, this Pixel 2 has 16Gb of RAM.

1 0

Re: [SeaBIOS] [HELP] The program cause screen mess
by eJim Lee 22 Apr '15

22 Apr '15

Hi kevin: If you want to repeat the bug, you can try: 1. Go to "Control panel->Regional and Language Options->Advanced->Language for non-Unicode programs" and select "Chinese (PRC)". 2. reboot 3. run wbc I also tried the program on vmware and looked nice. But something different with qemu about cursor: On qemu with WinXP English: cursor is on screen but not on vmware. Hope these information can help you. Thanks! eJim Lee 2015-04-21 23:41 GMT+08:00 eJim Lee <lyjforjob(a)gmail.com>: > I have tried it without kvm and bug repeated: > ./build/x86_64-softmmu/qemu-system-x86_64 -chardev stdio,id=seabios > -device isa-debugcon,iobase=0x402,chardev=seabios -m 512 -hda > tmp/xp.append -hdb tmp/data.append > > 2015-04-21 2:49 GMT+08:00 Kevin O'Connor <kevin(a)koconnor.net>: >> On Mon, Apr 20, 2015 at 02:46:03PM -0400, Kevin O'Connor wrote: >>> On Tue, Apr 21, 2015 at 01:57:53AM +0800, eJim Lee wrote: >>> > 2015-04-21 0:45 GMT+08:00 Kevin O'Connor <kevin(a)koconnor.net>: >>> > > On Tue, Apr 21, 2015 at 12:23:15AM +0800, eJim Lee wrote: >>> > >> Kevin, did you repeat screen mess? >>> > > >>> > > I've tried several different configurations (including the last >>> > > command line you sent), but I have not been able to reproduce any >>> > > screen corruption. >>> > >>> > I have repeated it at every time on server, my laptop or my pc. >>> > I tried it on ubuntu 14.04, centos 6.5, kernel 3.10.67 and kernel.3.16.0.36 etc. >>> > What is the difference? I used guest os is xp sp3. What yours? >>> > I will try again with environment same as yours. >>> >>> I'm using "Windows XP professional - Service Pack 3". >>> >>> To simplify versions, can you use the latest QEMU - 2.3.0-rc4 and use >>> the seabios and seavgabios that come with that version of qemu. I >>> tested that here and don't see corruption with those versions. This >>> is the command line I used: >>> >>> ../qemu/qemu-git/x86_64-softmmu/qemu-system-x86_64 -chardev stdio,id=seabios -device isa-debugcon,iobase=0x402,chardev=seabios -m 512 -hda winxp-test-wbc -enable-kvm >> >> Also, for completeness, try it without kvm as well: >> >> ../qemu/qemu-git/x86_64-softmmu/qemu-system-x86_64 -chardev stdio,id=seabios -device isa-debugcon,iobase=0x402,chardev=seabios -m 512 -hda winxp-test-wbc >> >> -Kevin

1 0

Problems booting legacy OSes on Chromebook Pixel 2015
by Charles Howell 22 Apr '15

22 Apr '15

(Apologies if this is the wrong list, but I'm not sure where else to post.) I'm attempting to boot FreeBSD 10.1 (amd64) on a new Chromebook Pixel (via a USB CD) and it hangs as soon as the kernel is loaded into memory. I'd just assume it's a FreeBSD problem but I've also tried the latest versions of OpenBSD and NetBSD with the same results. In all three cases, the bootloader loads and I can interact with it, toggle options, read machine memory layout, enable safe mode, etc. but as soon as the kernel is loaded into memory and it tries to execute, the machine hangs hard and I have to hold down power for a few seconds to shut it off. This leads me to believe it's a problem with SeaBIOS and/or its support for this machine since I've booted BSDs on the Acer C720 with SeaBIOS just fine. I've also tried with a USB key with an installed OS already in case it was related to reading the CD, but with the same results. The bootloader comes up but no kernels will boot. These were all tested with SeaBIOS version -20150225_120401-build122-m2 as shipped with the device. Can anyone offer any help?

1 0

about [SeaBIOS PATCH] Try to hard-reboot on rerun of post even on emulators.
by Amos Kong 21 Apr '15

21 Apr '15

Hi Kevin, When I use old seabios in some stable linux release, some bootable devices (2 ide disks) would be lost when I try to restart guest by Ctrl+Alt+Delete during boot stage. Releated Red Hat BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1129549 I found an upstream commit [1] fixed this bug, but when I backport this patch to old seabios, guest will shutdown when I try to restart by Ctrl+Alt+Delete during boot stage. Kevin, can you help to describe that: "Unfortunately, kvm does not keep a pristine copy of the BIOS at 0xffff0000" It's a kvm (userspace, QEMU) bug? If it's a qemu-kvm bug, I should also fix this bz in old stable release. [1] =========================================== commit 244caf86f11f5f65d166d91704f64cb673167abc Author: Kevin O'Connor <kevin(a)koconnor.net> Date: Wed Sep 15 21:48:16 2010 -0400 Try to hard-reboot on rerun of post even on emulators. Extend the hard-reboot logic to qemu and kvm. On qemu, a reboot will not reset the memory settings for 0xc0000-0xfffff, so copy that memory area manually before rebooting. Unfortunately, kvm does not keep a pristine copy of the BIOS at 0xffff0000, so detect that case and shutdown the machine. Two backport dependences: [PATCH] Try to hard-reboot processor on rerun of post under coreboot. [PATCH] Don't do shadow copying of optionroms when CONFIG_OPTIONROMS_DEPLOYED. -- Amos.

4 5

Re: [SeaBIOS] [HELP] The program cause screen mess
by Kevin O'Connor 21 Apr '15

21 Apr '15

On Tue, Apr 21, 2015 at 01:57:53AM +0800, eJim Lee wrote: > 2015-04-21 0:45 GMT+08:00 Kevin O'Connor <kevin(a)koconnor.net>: > > On Tue, Apr 21, 2015 at 12:23:15AM +0800, eJim Lee wrote: > >> Kevin, did you repeat screen mess? > > > > I've tried several different configurations (including the last > > command line you sent), but I have not been able to reproduce any > > screen corruption. > > I have repeated it at every time on server, my laptop or my pc. > I tried it on ubuntu 14.04, centos 6.5, kernel 3.10.67 and kernel.3.16.0.36 etc. > What is the difference? I used guest os is xp sp3. What yours? > I will try again with environment same as yours. I'm using "Windows XP professional - Service Pack 3". To simplify versions, can you use the latest QEMU - 2.3.0-rc4 and use the seabios and seavgabios that come with that version of qemu. I tested that here and don't see corruption with those versions. This is the command line I used: ../qemu/qemu-git/x86_64-softmmu/qemu-system-x86_64 -chardev stdio,id=seabios -device isa-debugcon,iobase=0x402,chardev=seabios -m 512 -hda winxp-test-wbc -enable-kvm -Kevin

2 3

回复:回复：Re: 回复：Re: _about ZoneHigh occupying ram from last address of below 4G ram
by penghao122＠sina.com 21 Apr '15

21 Apr '15

I just notice sending mail fail report ----- 原始邮件 ----- 发件人: penghao122<penghao122(a)sina.com> 收件人: Kevin O'Connor<kevin(a)koconnor.net> 抄送人: seabios<seabios(a)seabios.org> 主题: 回复：Re: 回复：Re: [SeaBIOS]_about ZoneHigh occupying ram from last address of below 4G ram 日期: 2015-4-18 00:28 my qemu command line: ../qemu-2.1.2/x86_64-softmmu/qemu-system-x86_64 -k en-us -chardev stdio,id=seabios -device isa-debugcon,iobase=0x402,chardev=seabios -smp 4 -m 8192 -drive file=/home/win7-64.qcow2,format=qcow2 -enable-kvm Just for linux OS ,I agree with you about "It's common to put reserved memory at the end of ram ". My win7-64 vm delay two hours before booting up successfully .During the two hours,the windows OS access from 0xc000_0000 to 0xfc00_0000 byte by byte .There is no pci devices in the address range. KVM think address space that is not in E820 table is mmio,but in fact this address space range also contain address hole. these hole should not be accessed by OS normally. Now windows OS access the hole ,so kvm think the address space is mmio,and handle the access like mmio flow (repeated kvm_exit to qemu userspace ).That is very time-consuming,so the vm delay 2 hours. I modified the ZoneHigh address at the begining of 128MB,everything is OK. The bootloader of windows os may be care about memory at the end of ram below 4GB. I try to start the same win7-64bit vm in xen-4.1.0(using rombios).It's OK because the rombios keep 0x10000 - 0xc000_0000 as ram.. p.s some edition of win7-64bit is easy to trigge ----- 原始邮件 ----- 发件人：Kevin O'Connor <kevin(a)koconnor.net> 收件人：penghao122(a)sina.com 抄送人：seabios <seabios(a)seabios.org> 主题：Re: 回复：Re: [SeaBIOS]_about ZoneHigh occupying ram from last address of below 4G ram 日期：2015年04月16日 22点22分 On Thu, Apr 16, 2015 at 10:08:48PM +0800, penghao122(a)sina.com wrote: > I configure ram to 4GB,then the win7-64bit vm can boot successfully > using unmodified seabios. And I find that the vm can boot up after > about 2 hours when configuring ram to 8G using unmodfied seabios. What is your qemu/kvm command line? It works fine for me (with a win7 beta I have locally). I do see a brief pause around the point where you show a 2 hour delay, but the delay for me is only a few seconds. This is the command I used: ../qemu/qemu-2.2.0/x86_64-softmmu/qemu-system-x86_64 -k en-us -snapshot -L test -chardev stdio,id=seabios -device isa-debugcon,iobase=0x402,chardev=seabios -m 8192 -hda win7beta.img -enable-kvm -vga cirrus -smp 2 -usb I think your issue should be reported on the QEMU/KVM mailing lists. It's common to put reserved memory at the end of ram - for example, run "dmesg | grep e820" on a real machine and see what it reports. So, this doesn't look like a seabios issue to me. -Kevin 来自新浪邮箱触屏版

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

SeaBIOS