SeaBIOS

seabios@seabios.org

3783 discussions

Re: [SeaBIOS] S3 resume fails on hp/8470p when using SeaBIOS 1.11.0
by Kevin O'Connor 25 Dec '17

25 Dec '17

Resending this mail as it looks like I accidently dropped the mailing list. On Tue, Dec 19, 2017 at 02:25:41PM +0800, Iru Cai wrote: > Hi, > > I'm using coreboot with SeaBIOS payload on hp/8470p. I found S3 resume > fails when using the latest SeaBIOS. After testing some revisions of > SeaBIOS, I found this commit is the problem. > > commit 04db972290a6f3c61920fc7498031f477911774e > Author: Kevin O'Connor <kevin(a)koconnor.net> > Date: Tue Jul 11 12:09:57 2017 -0400 > > disk: Don't require the 'struct drive_s' to be in the f-segment > > > From the cbmem console, when after using SeaBIOS>=commit 04db972, coreboot > cannot find RSDP, thus cannot do the ACPI wakeup. That's surprising. Can you provide the cbmem log from the success case and the failure case? -Kevin

1 0

Re: [SeaBIOS] cbfstool without cloning the whole repository? Use script for <2M download
by Ivan Ivanov 23 Dec '17

23 Dec '17

It looks like there were two typos in this wonderful script : missing " \ " at the end of two " --execute robots=off " lines Other than that, this script is still working. Below is a fixed copy, if you sometimes need cbfstool without cloning the whole coreboot you may try it out: ### SCRIPT BEGIN ### wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ --user-agent="Mozilla/5.0" --execute robots=off \ https://review.coreboot.org/cgit/coreboot.git/plain/util/cbfstool/ wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ --user-agent="Mozilla/5.0" --execute robots=off \ https://review.coreboot.org/cgit/coreboot.git/plain/src/commonlib/ \ --exclude-directories=/cgit/coreboot.git/plain/src/commonlib/storage/ \ --reject=configstring.c,iobuf.c,cbmem_id.h,configstring.h,coreboot_tables.h,\ fmap_serialized.h,iobuf.h,sd_mmc_ctrlr.h,sdhci.h,stdlib.h,storage.h,\ timestamp_serialized.h wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ --user-agent="Mozilla/5.0" --execute robots=off \ https://review.coreboot.org/cgit/coreboot.git/plain/src/vendorcode/intel/ed… wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ --user-agent="Mozilla/5.0" --execute robots=off \ https://review.coreboot.org/cgit/coreboot.git/plain/src/vendorcode/intel/fs… wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ --user-agent="Mozilla/5.0" --execute robots=off \ --directory-prefix=./3rdparty/vboot/ \ https://review.coreboot.org/cgit/vboot.git/plain/firmware/2lib/ \ --reject=2api.c,2crc8.c,2hmac.c,2misc.c,2nvstorage.c,2rsa.c,2secdata.c,\ 2secdatak.c,2stub.c,2tpm_bootmode.c,2crc8.h,2hmac.h,2misc.h,\ 2nvstorage_fields.h,2nvstorage.h,2rsa.h,2secdata.h,2tpm_bootmode.h wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ --user-agent="Mozilla/5.0" --execute robots=off \ --directory-prefix=./3rdparty/vboot/ \ https://review.coreboot.org/cgit/vboot.git/plain/firmware/include/vb2_api.h cd ./util/cbfstool/ make all ### SCRIPT END ### 2017-07-12 1:53 GMT+03:00 Ivan Ivanov <qmastery16(a)gmail.com>: > Dear friends, sometimes you need to do a few operations with > coreboot.rom (e.g. update its' payload) without changing > coreboot.rom's main code. If you have bad Internet connection or just > want to quickly get the latest version of cbfstool without cloning the > whole coreboot/vboot repositories - my bash script below will really > help you! > > It downloads only the files required for the successful compilation of > cbfstool - whole ./util/cbfstool directory (93 files, ~1,3M size) and > all the dependencies: > ./src/commonlib - 20 files, ~152K > ./src/vendorcode/intel/edk2/uefi_2.4/ - 26 files, 272K > ./src/vendorcode/intel/fsp/fsp1_1/IntelFspPkg/Include/FspInfoHeader.h - ~4K > > ((( I don't understand what Intel is doing here, but I can't build > without these files! ))) > > ./3rdparty/vboot/firmware/2lib/ - 17 files, 170K > ./3rdparty/vboot/firmware/include/vb2_api.h - ~1K > > In total, its 158 files and just 1.9M - less than 2M ! > And it builds OK. Hope it will be useful for you ;-) > > ### SCRIPT BEGIN ### > > wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ > --user-agent="Mozilla/5.0" --execute robots=off \ > https://review.coreboot.org/cgit/coreboot.git/plain/util/cbfstool/ > > wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ > --user-agent="Mozilla/5.0" --execute robots=off \ > https://review.coreboot.org/cgit/coreboot.git/plain/src/commonlib/ \ > --exclude-directories=/cgit/coreboot.git/plain/src/commonlib/storage/ \ > --reject=configstring.c,iobuf.c,cbmem_id.h,configstring.h,coreboot_tables.h,\ > fmap_serialized.h,iobuf.h,sd_mmc_ctrlr.h,sdhci.h,stdlib.h,storage.h,\ > timestamp_serialized.h > > wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ > --user-agent="Mozilla/5.0" --execute robots=off \ > https://review.coreboot.org/cgit/coreboot.git/plain/src/vendorcode/intel/ed… > > wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ > --user-agent="Mozilla/5.0" --execute robots=off \ > https://review.coreboot.org/cgit/coreboot.git/plain/src/vendorcode/intel/fs… > > wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ > --user-agent="Mozilla/5.0" --execute robots=off > --directory-prefix=./3rdparty/vboot/ \ > https://review.coreboot.org/cgit/vboot.git/plain/firmware/2lib/ \ > --reject=2api.c,2crc8.c,2hmac.c,2misc.c,2nvstorage.c,2rsa.c,2secdata.c,\ > 2secdatak.c,2stub.c,2tpm_bootmode.c,2crc8.h,2hmac.h,2misc.h,\ > 2nvstorage_fields.h,2nvstorage.h,2rsa.h,2secdata.h,2tpm_bootmode.h > > wget --recursive --no-parent --no-host-directories --cut-dirs=3 \ > --user-agent="Mozilla/5.0" --execute robots=off > --directory-prefix=./3rdparty/vboot/ \ > https://review.coreboot.org/cgit/vboot.git/plain/firmware/include/vb2_api.h > > cd ./util/cbfstool/ > > make all > > ### SCRIPT END ###

1 0

[PATCH 0/1] add multiple usb-hid support
by Stef van Os 23 Dec '17

23 Dec '17

The following patch adds support for multiple usb keyboards and mice to seabios. Maximum number of keyboards is configurable in Kconfig and defaults to 1 to keep current behavior. Tested on: * Qemu q35 with two usb keyboards on seperate displays * Prodrive broadwell-d platform with one Dell keyboard and one virtual KVM keyboard on Aspeed AST2500. Stef van Os (1): usb-hid: add support for multiple input devices src/Kconfig | 14 +++++++++++ src/hw/usb-hid.c | 76 ++++++++++++++++++++++++++++++++++++++++---------------- 2 files changed, 68 insertions(+), 22 deletions(-) -- 2.11.0

3 5

Re: [SeaBIOS] [PATCH 0/1] add multiple usb-hid support
by Stef van Os 22 Dec '17

22 Dec '17

Hi Daniel, Is this a problem with the patch applied? It was tested on our hardware to function. Some questions: * What hardware platform are you using? * Does the numeric keyboard work if it is the only keyboard (or if you swap ports, which hopefully changes enumeration order) in the system, maybe the problem is elsewhere and was masked because the (normal) keyboard was enumerated first and the second (numeric) was ignored earlier. Also, the patch can be verified using qemu and two vnc inputs: qemu-system-x86_64 -m 4096 -M q35 -smp 8 \ -pflash bios.qemu.rom \ -drive format=raw,file=disk.img \ -device pci-bridge,addr=12.0,chassis_nr=2,id=head.2 \ -device nec-usb-xhci,bus=head.2,addr=0f.0,id=usb.2 \ -device VGA,id=video.1 \ -device secondary-vga,bus=head.2,addr=02.0,id=video.2 \ -display vnc=:15,id=primary,display=video.1 \ -display vnc=:16,id=secondary,display=video.2 \ -usb \ -device usb-kbd,bus=usb.2.0,port=1,display=video.1 \ -device usb-kbd,bus=usb.2.0,port=2,display=video.2 \ -serial stdio Best regards, Stef On 12/20/2017 06:20 PM, Daniel K wrote: > When I have more than 2 usb keyboards connected, SeaBIOS doesn't register keyboard input. > This persists even when I get to the disk GRUB menu. > I have to manually unplug my other keyboard to do any input during boot. > > And by other keyboard, I just mean a separate numerical keypad. > >> From: SeaBIOS <seabios-bounces(a)seabios.org> on behalf of Kevin O'Connor <kevin(a)koconnor.net> >> Sent: 20 December 2017 08:33 >> To: Stef van Os >> Cc: seabios(a)seabios.org >> Subject: Re: [SeaBIOS] [PATCH 0/1] add multiple usb-hid support >> >> On Fri, Dec 15, 2017 at 11:09:45AM +0100, Stef van Os wrote: >>> The following patch adds support for multiple usb keyboards and mice to seabios. >>> Maximum number of keyboards is configurable in Kconfig and defaults to 1 to keep >>> current behavior. >> >> Thanks. What's the use case for enabling multiple keyboards? >> >> -Kevin > > > >

2 1

Re: [SeaBIOS] [PATCH 0/1] add multiple usb-hid support
by Daniel K 21 Dec '17

21 Dec '17

Sorry I meant more than 1 usb keyboard. So 2 or more will cause the issue. >From: SeaBIOS <seabios-bounces(a)seabios.org> on behalf of Daniel K <shambles_09(a)hotmail.com> >Sent: 20 December 2017 09:20 >To: Kevin O'Connor; Stef van Os >Cc: seabios(a)seabios.org >Subject: Re: [SeaBIOS] [PATCH 0/1] add multiple usb-hid support > > >When I have more than 2 usb keyboards connected, SeaBIOS doesn't register keyboard input. >This persists even when I get to the disk GRUB menu. >I have to manually unplug my other keyboard to do any input during boot. > >And by other keyboard, I just mean a separate numerical keypad. > >>From: SeaBIOS <seabios-bounces(a)seabios.org> on behalf of Kevin O'Connor <kevin(a)koconnor.net> >>Sent: 20 December 2017 08:33 >>To: Stef van Os >>Cc: seabios(a)seabios.org >>Subject: Re: [SeaBIOS] [PATCH 0/1] add multiple usb-hid support >> >>On Fri, Dec 15, 2017 at 11:09:45AM +0100, Stef van Os wrote: >>> The following patch adds support for multiple usb keyboards and mice to seabios. >>> Maximum number of keyboards is configurable in Kconfig and defaults to 1 to keep >>> current behavior. >> >> Thanks. What's the use case for enabling multiple keyboards? >> >> -Kevin

1 0

S3 resume fails on hp/8470p when using SeaBIOS 1.11.0
by Iru Cai 19 Dec '17

19 Dec '17

Hi, I'm using coreboot with SeaBIOS payload on hp/8470p. I found S3 resume fails when using the latest SeaBIOS. After testing some revisions of SeaBIOS, I found this commit is the problem. commit 04db972290a6f3c61920fc7498031f477911774e Author: Kevin O'Connor <kevin(a)koconnor.net> Date: Tue Jul 11 12:09:57 2017 -0400 disk: Don't require the 'struct drive_s' to be in the f-segment >From the cbmem console, when after using SeaBIOS>=commit 04db972, coreboot cannot find RSDP, thus cannot do the ACPI wakeup. Iru -- Please do not send me Microsoft Office/Apple iWork documents. Send OpenDocument instead! http://fsf.org/campaigns/opendocument/

1 0

[PATCH] readserial: allow setting real time timestamps
by Vadim Bendebury (вб) 19 Dec '17

19 Dec '17

Sometimes it is beneficial to attach real time stamps to console output, for instance when there is a need to correlate two independently captured logs. This patch adds another command line option (-r,--realtime), when given the generated log is printed with second's resolution real time timestamps. Signed-off-by: Vadim Bendebury <vbendeb(a)chromium.org> --- scripts/readserial.py | 44 +++++++++++++++++++++++++++----------------- 1 file changed, 27 insertions(+), 17 deletions(-) diff --git a/scripts/readserial.py b/scripts/readserial.py index a7383e8..dc294eb 100755 --- a/scripts/readserial.py +++ b/scripts/readserial.py @@ -8,7 +8,7 @@ # Usage: # scripts/readserial.py /dev/ttyUSB0 115200 -import sys, os, time, select, optparse +import datetime, sys, os, time, select, optparse from python23compat import as_bytes @@ -54,8 +54,10 @@ def calibrateserialread(infile, byteadjust): totalchars += len(d) lasttime = curtime -def readserial(infile, logfile, byteadjust): +def readserial(infile, logfile, byteadjust, real_time): lasttime = 0 + isnewline = 0 + charcount = 0 while 1: # Read data try: @@ -74,25 +76,29 @@ def readserial(infile, logfile, byteadjust): return 0 datatime = time.time() - datatime -= len(d) * byteadjust + if not real_time: + datatime -= len(d) * byteadjust - # Reset start time if no data for some time - if datatime - lasttime > RESTARTINTERVAL: - starttime = datatime - charcount = 0 - isnewline = 1 - msg = "\n\n======= %s (adjust=%.1fus)\n" % ( - time.asctime(time.localtime(datatime)), byteadjust * 1000000) - sys.stdout.write(msg) - logfile.write(as_bytes(msg)) - lasttime = datatime + # Reset start time if no data for some time + if datatime - lasttime > RESTARTINTERVAL: + starttime = datatime + charcount = 0 + isnewline = 1 + msg = "\n\n======= %s (adjust=%.1fus)\n" % ( + time.asctime(time.localtime(datatime)), byteadjust * 1000000) + sys.stdout.write(msg) + logfile.write(as_bytes(msg)) + lasttime = datatime # Translate unprintable chars; add timestamps out = as_bytes("") for c in d: if isnewline: - delta = datatime - starttime - (charcount * byteadjust) - out += "%06.3f: " % delta + if real_time: + out += datetime.datetime.now().strftime('%H:%M:%S: ') + else: + delta = datatime - starttime - (charcount * byteadjust) + out += "%06.3f: " % delta isnewline = 0 oc = ord(c) charcount += 1 @@ -138,6 +144,10 @@ def main(): opts.add_option("-t", "--time", type="float", dest="time", default=None, help="time to write one byte on serial port (in us)") + opts.add_option("-r", "--realtime", + action="store_true", dest="realtime", default=False, + help="print real time timestamps instead of deltas") + options, args = opts.parse_args() serialport = 0 baud = 115200 @@ -176,12 +186,12 @@ Or: apt-get install python-serial logname = time.strftime("seriallog-%Y%m%d_%H%M%S.log") f = open(logname, 'wb') if options.serial: - readserial(ser, f, byteadjust) + readserial(ser, f, byteadjust, options.realtime) else: # Read from a pipe while 1: ser = os.fdopen(os.open(serialport, os.O_RDONLY|os.O_NONBLOCK), 'rb') - res = readserial(ser, f, byteadjust) + res = readserial(ser, f, byteadjust, options.realtime) ser.close() if res < 0: break -- 2.15.1.504.g5279b80103-goog

1 0

[PATCH] readserial: allow reading of zero bytes
by Vadim Bendebury 19 Dec '17

19 Dec '17

Certain serial device drivers might return without reading anything, returning empty buffer. This is not an indication of the interface being shut down, the readserial utility should keep running. Signed-off-by: Vadim Bendebury <vbendeb(a)chromium.org> --- scripts/readserial.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/readserial.py b/scripts/readserial.py index a7383e8..ffb076f 100755 --- a/scripts/readserial.py +++ b/scripts/readserial.py @@ -71,7 +71,7 @@ def readserial(infile, logfile, byteadjust): continue d = infile.read(4096) if not d: - return 0 + continue datatime = time.time() datatime -= len(d) * byteadjust -- 2.15.1.504.g5279b80103-goog

1 0

Re: [SeaBIOS] Loading Option ROM for SATA Controller Card
by Kevin O'Connor 10 Dec '17

10 Dec '17

On Sun, Nov 26, 2017 at 06:28:01PM -0800, Adam Novak wrote: > I've installed a SI-PEX40064 SATA controller card > <http://www.sybausa.com/index.php?route=product/product&product_id=156> > in my system that I am trying to use with PCI passthrough via Xen, in > order to give a guest VM its own real SATA controller so as to not > have to deal with virtual disks. Moreover, I want to boot off of a > disk attached to the card. > > The card implements (or pretends to implement) AHCI, so SeaBIOS starts > up and can see the attached disk. But SeaBIOS gets a timeout > communicating with the disk: > > |eff9d000| WARNING - Timeout at ahci_port_setup:469! > |eff9d000| AHCI/2: device not ready (tf 0x58) > > The card has a PCI Option ROM, and I think the Option ROM needs to be > loaded for the card to work and/or be bootable. But SeaBIOS refuses to > load the Option ROM. (My real physical system with Asus EFI firmware > and CSM enabled does load the Option ROM for the card when it is > installed.) I dug into the code, and it looks like it's the same issue > as is mentioned here: > > https://mail.coreboot.org/pipermail/seabios/2017-June/011425.html > > Basically, a device is not checked for an option ROM (or an > fw_fcg/CBFS-based Option ROM file), and its option ROM is not loaded, > if pci->have_driver is set for the device. That field gets set for > anything that apparently implements AHCI, so the Option ROMs for AHCI > controllers are skipped, even if they are useful. Removing that check > causes the option ROM for my card to be loaded, and allows the > bootloader stored on the disk attached to the card to start up. That check is there because it would not be correct for both an option rom and a native driver to simultaneously attempt to register the same drive. Basically, there can be only one driver for a PCI card - either a native SeaBIOS one or an Option ROM. What happens if you compile SeaBIOS with CONFIG_AHCI disabled? -Kevin

2 1

[PATCH] docs/Download: Use more secure HTTPS ULRs where possible
by Paul Menzel 09 Dec '17

09 Dec '17

Date: Mon, 2 Oct 2017 08:13:13 +0200 Signed-off-by: Paul Menzel <pmenzel(a)molgen.mpg.de> --- docs/Download.md | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/docs/Download.md b/docs/Download.md index 9b1492a..96369fb 100644 --- a/docs/Download.md +++ b/docs/Download.md @@ -1,11 +1,11 @@ SeaBIOS may be distributed under the terms of the [GNU -LGPLv3](http://www.gnu.org/licenses/lgpl-3.0-standalone.html) license. +LGPLv3](https://www.gnu.org/licenses/lgpl-3.0-standalone.html) license. Both source code and binaries are available. Latest source code ================== -The SeaBIOS project uses the [git](http://git-scm.com/) revision +The SeaBIOS project uses the [git](https://git-scm.com/) revision control system. To download the latest source from revision control, run: @@ -22,6 +22,6 @@ Released versions Released versions of the source code are available at: -<http://code.coreboot.org/p/seabios/downloads/> +<https://code.coreboot.org/p/seabios/downloads/> Please see [releases](Releases) for information on each release. -- 2.14.1

2 1

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

SeaBIOS