SeaBIOS

seabios@seabios.org

3783 discussions

Issues with AHCI and AMD SB700
by Paul Menzel 13 Nov '15

13 Nov '15

Dear SeaBIOS folks, Timothy Pearson submitted the change below [1] to the coreboot repository. ``` ommit 14ebf951e41c085aa2f8420fcf7452dc5761514e Author: Timothy Pearson <tpearson(a)raptorengineeringinc.com> Date: Sun Jun 21 16:27:03 2015 -0500 src/southbridge/amd/sb700: Reset SATA controller in AHCI mode during startup In AHCI mode SeaBIOS randomly fails to detect disks (AHCI timeouts), with the probability of a failure increasing with the number of disks connected to the controller. Resetting the SATA controller appears to show the true state of the underlying hardware, allowing the drive detection code to attempt link renegotiation as needed. Change-Id: Ib1f7c5f830a0cdba41cb6f5b05d759adee5ce369 Signed-off-by: Timothy Pearson <tpearson(a)raptorengineeringinc.com> Reviewed-on: http://review.coreboot.org/11998 Tested-by: Raptor Engineering Automated Test Stand <noreply(a)raptorengineeringinc.com> Reviewed-by: Stefan Reinauer <stefan.reinauer(a)coreboot.org> Tested-by: build bot (Jenkins) diff --git a/src/southbridge/amd/sb700/sata.c b/src/southbridge/amd/sb700/sata.c index ae44446..63c07d8 100644 --- a/src/southbridge/amd/sb700/sata.c +++ b/src/southbridge/amd/sb700/sata.c @@ -131,6 +131,8 @@ static void sata_init(struct device *dev) /* get rev_id */ rev_id = pci_read_config8(sm_dev, 0x08) - 0x28; + printk(BIOS_SPEW, "rev_id=%x\n", rev_id); + if (sata_ahci_mode) { /* Enable link latency enhancement on A14 and above */ if (rev_id >= 0x14) { @@ -241,6 +243,27 @@ static void sata_init(struct device *dev) write32(sata_bar5 + 0xfc, dword); } + if (sata_ahci_mode) { + /* FIXME + * SeaBIOS does not know how to spin + * up the drives and therefore hangs + * in AHCI init if this is enabled... + */ + /* Enable staggered spin-up */ + dword = read32(sata_bar5 + 0x00); +#if 0 + dword |= 0x1 << 27; +#else + dword &= ~(0x1 << 27); +#endif + write32(sata_bar5 + 0x00, dword); + + /* Reset the HBA to avoid stuck drives in SeaBIOS */ + dword = read32(sata_bar5 + 0x04); + dword |= 0x1; + write32(sata_bar5 + 0x04, dword); + } + /* Write protect Sub-Class Code */ byte = pci_read_config8(dev, 0x40); byte &= ~(1 << 0); @@ -367,21 +390,21 @@ static void sata_init(struct device *dev) } /* Below is CIM InitSataLateFar */ - /* Enable interrupts from the HBA */ - byte = read8(sata_bar5 + 0x4); - byte |= 1 << 1; - write8((sata_bar5 + 0x4), byte); - if (!sata_ahci_mode) { - /* Clear error status */ - write32((sata_bar5 + 0x130), 0xFFFFFFFF); - write32((sata_bar5 + 0x1b0), 0xFFFFFFFF); - write32((sata_bar5 + 0x230), 0xFFFFFFFF); - write32((sata_bar5 + 0x2b0), 0xFFFFFFFF); - write32((sata_bar5 + 0x330), 0xFFFFFFFF); - write32((sata_bar5 + 0x3b0), 0xFFFFFFFF); + /* Enable interrupts from the HBA */ + byte = read8(sata_bar5 + 0x4); + byte |= 1 << 1; + write8((sata_bar5 + 0x4), byte); } + /* Clear error status */ + write32((sata_bar5 + 0x130), 0xFFFFFFFF); + write32((sata_bar5 + 0x1b0), 0xFFFFFFFF); + write32((sata_bar5 + 0x230), 0xFFFFFFFF); + write32((sata_bar5 + 0x2b0), 0xFFFFFFFF); + write32((sata_bar5 + 0x330), 0xFFFFFFFF); + write32((sata_bar5 + 0x3b0), 0xFFFFFFFF); + /* Clear SATA status,Firstly we get the AcpiGpe0BlkAddr */ /* ????? why CIM does not set the AcpiGpe0BlkAddr , but use it??? */ /* ????? why CIM does not set the AcpiGpe0BlkAddr , but use it??? */ ``` Do you know if there is a fix for SeaBIOS or is this a hardware issue? Thanks, Paul [1] http://review.coreboot.org/11998

1 0

[PATCH] buildversion: Avoid subprocess.check_output() as that requires python2.7
by Kevin O'Connor 12 Nov '15

12 Nov '15

Don't require python2.7 in buildversion.py. Also, ignore only those exceptions that are known to be possible. Signed-off-by: Kevin O'Connor <kevin(a)koconnor.net> --- scripts/buildversion.py | 30 +++++++++++++++++------------- 1 file changed, 17 insertions(+), 13 deletions(-) diff --git a/scripts/buildversion.py b/scripts/buildversion.py index d61fc9e..5b8807e 100755 --- a/scripts/buildversion.py +++ b/scripts/buildversion.py @@ -4,7 +4,7 @@ # Copyright (C) 2015 Kevin O'Connor <kevin(a)koconnor.net> # # This file may be distributed under the terms of the GNU GPLv3 license. -import sys, os, subprocess, time, socket, optparse +import sys, os, subprocess, shlex, time, socket, optparse VERSION_FORMAT = """ /* DO NOT EDIT! This is an autogenerated file. See scripts/buildversion.py. */ @@ -12,16 +12,24 @@ VERSION_FORMAT = """ #define BUILD_TOOLS "%s" """ +# Run program and return the specified output +def check_output(prog): + try: + process = subprocess.Popen(shlex.split(prog), stdout=subprocess.PIPE) + output = process.communicate()[0] + retcode = process.poll() + except OSError: + return "" + try: + return output.decode() + except UnicodeError: + return "" + # Obtain version info from "git" program def git_version(): if not os.path.exists('.git'): return "" - params = "git describe --tags --long --dirty".split() - try: - ver = subprocess.check_output(params).decode().strip() - except: - return "" - return ver + return check_output("git describe --tags --long --dirty").strip() # Look for version in a ".version" file. Official release tarballs # have this file (see scripts/tarball.sh). @@ -32,7 +40,7 @@ def file_version(): f = open('.version', 'r') ver = f.readline().strip() f.close() - except: + except OSError: return "" return ver @@ -50,11 +58,7 @@ def tool_versions(tools): success = 0 for tool in tools: # Extract first line from "tool --version" output - try: - ver = subprocess.check_output([tool, '--version']).decode() - except: - continue - verstr = ver.split('\n')[0] + verstr = check_output("%s --version" % (tool,)).split('\n')[0] # Check if this tool looks like a binutils program isbinutils = 0 if verstr.startswith('GNU '): -- 2.4.3

3 5

WD elements problem
by Wim Vervoorn 11 Nov '15

11 Nov '15

Hello Kevin, I hooked up an external USB 3.0 WD hard drive and found that this wasn't supported by SeaBIOS as this used 4K blocks. Do you have plans to support this in the future in some way? Or will this simply be incompatible without plans for support. (This isn't a problem but it's good to know what the plans are). I haven't seen any devices like that - I was under the impression the 4K drives emulated 512 byte blocks. It should be possible to support them though. Can you send an email to the mailing list (and include a debug log)? BR, Wim

2 7

[PATCH] buildversion: Add debugging messages
by Kevin O'Connor 11 Nov '15

11 Nov '15

Add ability to output debug messages from the buildversion.py build script. Signed-off-by: Kevin O'Connor <kevin(a)koconnor.net> --- scripts/buildversion.py | 21 +++++++++++++++++++-- 1 file changed, 19 insertions(+), 2 deletions(-) diff --git a/scripts/buildversion.py b/scripts/buildversion.py index 1045049..ad47b90 100755 --- a/scripts/buildversion.py +++ b/scripts/buildversion.py @@ -4,7 +4,7 @@ # Copyright (C) 2015 Kevin O'Connor <kevin(a)koconnor.net> # # This file may be distributed under the terms of the GNU GPLv3 license. -import sys, os, subprocess, shlex, time, socket, optparse +import sys, os, subprocess, shlex, time, socket, optparse, logging, traceback VERSION_FORMAT = """ /* DO NOT EDIT! This is an autogenerated file. See scripts/buildversion.py. */ @@ -14,40 +14,51 @@ VERSION_FORMAT = """ # Run program and return the specified output def check_output(prog): + logging.debug("Running %s" % (repr(prog),)) try: process = subprocess.Popen(shlex.split(prog), stdout=subprocess.PIPE) output = process.communicate()[0] retcode = process.poll() except OSError: + logging.debug("Exception on run: %s" % (traceback.format_exc(),)) return "" + logging.debug("Got (code=%s): %s" % (retcode, repr(output))) if retcode: return "" try: return output.decode() except UnicodeError: + logging.debug("Exception on decode: %s" % (traceback.format_exc(),)) return "" # Obtain version info from "git" program def git_version(): if not os.path.exists('.git'): + logging.debug("No '.git' file/directory found") return "" - return check_output("git describe --tags --long --dirty").strip() + ver = check_output("git describe --tags --long --dirty").strip() + logging.debug("Got git version: %s" % (repr(ver),)) + return ver # Look for version in a ".version" file. Official release tarballs # have this file (see scripts/tarball.sh). def file_version(): if not os.path.isfile('.version'): + logging.debug("No '.version' file found") return "" try: f = open('.version', 'r') ver = f.readline().strip() f.close() except OSError: + logging.debug("Exception on read: %s" % (traceback.format_exc(),)) return "" + logging.debug("Got .version: %s" % (repr(ver),)) return ver # Generate an output file with the version information def write_version(outfile, version, toolstr): + logging.debug("Write file %s and %s" % (repr(version), repr(toolstr))) sys.stdout.write("Version: %s\n" % (version,)) f = open(outfile, 'w') f.write(VERSION_FORMAT % (version, toolstr)) @@ -74,6 +85,8 @@ def tool_versions(tools): continue # Check for any version conflicts if versions[isbinutils] and versions[isbinutils] != ver: + logging.debug("Mixed version %s vs %s" % ( + repr(versions[isbinutils]), repr(ver))) vers[isbinutils] = "mixed" continue versions[isbinutils] = ver @@ -88,11 +101,15 @@ def main(): help="extra version string to append to version") opts.add_option("-t", "--tools", dest="tools", default="", help="list of build programs to extract version from") + opts.add_option("-v", action="store_true", dest="verbose", + help="enable debug messages") options, args = opts.parse_args() if len(args) != 1: opts.error("Incorrect arguments") outfile = args[0] + if options.verbose: + logging.basicConfig(level=logging.DEBUG) cleanbuild, toolstr = tool_versions(options.tools) -- 2.4.3

1 0

Delay next release
by Kevin O'Connor 10 Nov '15

10 Nov '15

I was originally planning for a 1.9.0 release on Nov 6th. However, a couple of issues were raised this past week (python2.7 dependency in buildversion.py, USB3 failures, reboot race condition). As a result, I think we should push back the release a week (target Nov 16th). Here are my draft release notes for 1.9.0: * The default boot menu key is now the ESC key (instead of F12) * Initial support for Trusted Platform Module (TPM) hardware and BIOS calls * Initial support for chain loading SeaBIOS from Grub (via multiboot support) * virtio 1.0 device support * The build will no longer include the build hostname or build time on "clean" builds. This makes the build binaries more "reproducible". * SeaVGABIOS improvements: * Improved support for old versions of x86emu (the "leal" instruction is now emulated) * Several bug fixes and code cleanups Comments? -Kevin

1 1

[PATCH 0/3] Patches for baytrail based chromebooks
by Kevin O'Connor 09 Nov '15

09 Nov '15

This series enables seabios to run on some Baytrail CPU based chromebooks. At least some of these machines do not support routing of legacy interrupts and at least some have SDHCI controllers that do not appear as regular PCI devices. This series is mainly a hack to get some minimal support on the hardware. This series is also available at: https://github.com/KevinOConnor/seabios/tree/baytrail-testing -Kevin Kevin O'Connor (3): Add minimal support for machines without hardware interrupts ps2: Eliminate "etc/ps2-poll-only"; use CONFIG_HARDWARE_IRQ instead sdcard: Allow sdcard addresses to be specified in CBFS files docs/Runtime_config.md | 2 +- src/Kconfig | 13 ++++++++++++- src/clock.c | 31 +++++++++++++++++++++++++------ src/hw/pic.c | 14 ++++++++++++++ src/hw/pic.h | 4 ++++ src/hw/ps2port.c | 16 ++++++---------- src/hw/sdcard.c | 48 ++++++++++++++++++++++++++++++++++++------------ src/hw/timer.c | 2 ++ src/stacks.c | 5 ++++- src/util.h | 1 + 10 files changed, 105 insertions(+), 31 deletions(-) -- 1.9.3

3 14

Re: [SeaBIOS] [Qemu-devel][PATCH] SeaBios: Fix reset procedure reentrancy problem on qemu-kvm platform
by Gonglei 06 Nov '15

06 Nov '15

Ccing Seabios community. On 2015/11/3 14:58, Xulei (Stone, Euler) wrote: > On qemu-kvm platform, when I reset a VM through "virsh reset", and coincidently > the VM is in process of internal rebooting at the same time. Then the VM will > not be successfully reseted any more due to the reset reentrancy. I found: > (1)SeaBios try to shutdown the VM after reseting it failed by apm_shutdown(). > However, apm_shutdown() does not work on qemu-kvm platform; > (2)I add 1s sleep in qemu_prep_reset(), then continuously reset the VM twice, > aforementioned case must happen. > This patch fixes this issue by letting the VM always execute the reboot > routing while a reenrancy happenes instead of attempting apm_shutdown on > qemu-kvm platform. > > Signed-off-by: Lei Xu <stone.xulei(a)huawei.com> > --- > roms/seabios/src/resume.c | 7 ++++++- > 1 file changed, 6 insertions(+), 1 deletion(-) > > diff --git a/roms/seabios/src/resume.c b/roms/seabios/src/resume.c > index 1903174..96ff79e 100644 > --- a/roms/seabios/src/resume.c > +++ b/roms/seabios/src/resume.c > @@ -16,6 +16,7 @@ > #include "std/bda.h" // struct bios_data_area_s > #include "string.h" // memset > #include "util.h" // dma_setup > +#include "fw/paravirt.h" //runningOnKVM > > // Handler for post calls that look like a resume. > void VISIBLE16 > @@ -122,7 +123,11 @@ tryReboot(void) > dprintf(1, "Unable to hard-reboot machine - attempting shutdown.\n"); > apm_shutdown(); > } > - HaveAttemptedReboot = 1; > + if (!runningOnKVM()) { > + // Hard reboot has failed - try to shutdown machine. > + HaveAttemptedReboot = 1; > + } > + > > dprintf(1, "Attempting a hard reboot\n"); > > -- > 1.7.12.4 >

3 2

Re: [SeaBIOS] [PATCH] PCI SDHCI driver: Fix base address
by Kevin O'Connor 05 Nov '15

05 Nov '15

On Wed, Nov 04, 2015 at 07:40:12AM +0200, Kyösti Mälkki wrote: > MMIO register space was off by 4 bytes, bottom bits of > PCI BAR registers are not part of the address. Thanks! I committed this patch. -Kevin

1 0

[PATCH] PCI SDHCI driver: Fix base address
by Kyösti Mälkki 04 Nov '15

04 Nov '15

MMIO register space was off by 4 bytes, bottom bits of PCI BAR registers are not part of the address. Test boot on real hardware: PCI device 00:14.7 (vd=1022:7813 c=0805) Searching bootorder for: /pci@i0cf8/*@14,7 sdhci@0xfeb25500 ver=1001 cap=31fe3fb2 2570 sdcard_set_frequency 63 400 8000 sdcard_pio command stop (code=1) sdcard_set_frequency 63 25000 200 Found sdcard at 0xfeb25500: SD card SD01G 982MiB Registering bootable: SD card SD01G 982MiB (type:2 prio:103 data:f46e0) Signed-off-by: Kyösti Mälkki <kyosti.malkki(a)gmail.com> --- src/hw/sdcard.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/hw/sdcard.c b/src/hw/sdcard.c index d80b473..be726d2 100644 --- a/src/hw/sdcard.c +++ b/src/hw/sdcard.c @@ -527,6 +527,7 @@ sdcard_pci_setup(void *data) wait_preempt(); // Avoid pci_config_readl when preempting // XXX - bars dependent on slot index register in pci config space u32 regs = pci_config_readl(pci->bdf, PCI_BASE_ADDRESS_0); + regs &= PCI_BASE_ADDRESS_MEM_MASK; pci_config_maskw(pci->bdf, PCI_COMMAND, 0, PCI_COMMAND_IO | PCI_COMMAND_MEMORY | PCI_COMMAND_MASTER); int prio = bootprio_find_pci_device(pci); -- 1.9.1

1 0

Re: [SeaBIOS] [RFC PATCH] x86: Support booting SeaBIOS
by Bin Meng 02 Nov '15

02 Nov '15

Hi Simon, On Mon, Oct 19, 2015 at 10:51 AM, Simon Glass <sjg(a)chromium.org> wrote: > Hi Bin, > > On 10 October 2015 at 02:57, Bin Meng <bmeng.cn(a)gmail.com> wrote: >> Hi Simon, >> >> On Sat, Oct 3, 2015 at 10:29 PM, Simon Glass <sjg(a)chromium.org> wrote: >>> Hi Bin, >>> >>> On 29 September 2015 at 11:17, Bin Meng <bmeng.cn(a)gmail.com> wrote: >>>> SeaBIOS is an open source implementation of a 16-bit X86 BIOS. >>>> It can run in an emulator or natively on X86 hardware with the >>>> use of coreboot. With SeaBIOS's help, we can boot some OSes >>>> that require 16-bit BIOS services like Windows/DOS. >>>> >>>> As U-Boot, we have to manually create a table where SeaBIOS gets >>>> system information (eg: E820) from. The table unfortunately has >>>> to follow the coreboot table format as SeaBIOS currently supports >>>> booting as a coreboot payload. No U-Boot native support there. >>>> >>>> Booting SeaBIOS is done via U-Boot's bootelf command. >>>> >>>> This is the initial attempt to support booting SeaBIOS from U-Boot. >>>> If the basic concept is good, I can spend time working on follow-on >>>> patches to enable BIOS tables as well as graphics support. One issue >>>> is that U-Boot x86 does not has a ROM file system like coreboot. >>>> This brings difficulities to pass PCI option ROM to SeaBIOS, if we >>>> don't modify SeaBIOS's source codes. Maybe we should promote CBFS >>>> in U-Boot x86? >>>> >>>> This is tested on an Intel Crown Bay board with VGA card, booting >>>> SeaBIOS then chain loading a GRUB on a USB drive, then Linux kernel >>>> finally. >>> >>> Looks good to me. I think it is OK to use CBFS if needed - are you >>> thinking of an option to build u-boot.rom as a CBFS filesystem? >> >> If using CBFS, that means we may have to abandon ifdtool? Or maybe >> mixed usage of both tools? > > So far I'm not sure of the best approach. At present we have the ROM > offsets stored mostly in Kconfig, with the MRC area in the SPI flash > device tree node. The environment is also in Kconfig. > > What sort of option ROMs do you want to support? What other options > does seabios provide? > I believe SeaBIOS can run any kind of PCI option ROMs, but I have never tested other than vgabios. Also per coreboot's documentation, booting Windows needs SeaBIOS. > What does SEA stand for? I guess it's because most of SeaBIOS's source codes are written in 'C', thus a 'C' BIOS. Regards, Bin

1 0

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

SeaBIOS