coreboot December 2011

coreboot@coreboot.org

34 participants
174 discussions

[PATCH] kbd: wait longer for self-test on keyboard reset

by Mathias Krause

Some keyboards take pretty long to respond to a reset command, some even delay the ACK to the command. To make the keyboard driver more robust, increase the timeout for this special command. Also do an interface test after the self-test to ensure the keyboard is functioning properly. Another point is to reenable the keyboard *after* the scancode was set, not before. We also set the system bit when enabling the keyboard because this seems to be what older operating systems do expect. One of the problematic keyboards, which will work with this patch applied, is the DELL RT7D20. Without the patch an overly optimistic operating system, read Linux 2.4, will not recognise the keyboard because coreboot didn't fully initialize it. Signed-off-by: Mathias Krause <mathias.krause(a)secunet.com> --- Since I was unable to download the git tree (why is it HTTP only, btw?) a patch against a tarball snapshot. diff -urp coreboot-403d2d6/src/pc80/keyboard.c coreboot-kbd_fix/src/pc80/keyboard.c --- coreboot-403d2d6/src/pc80/keyboard.c 2011-12-08 22:42:04.000000000 +0100 +++ coreboot-kbd_fix/src/pc80/keyboard.c 2011-12-12 10:16:04.000000000 +0100 @@ -36,6 +36,7 @@ #define KBC_CMD_READ_COMMAND 0x20 // Read command byte #define KBC_CMD_WRITE_COMMAND 0x60 // Write command byte #define KBC_CMD_SELF_TEST 0xAA // Controller self-test +#define KBC_CMD_KBD_TEST 0xAB // Keyboard Interface test /* The Keyboard controller command byte * BIT | Description @@ -136,6 +137,26 @@ static int kbc_self_test(void) return 0; } + /* ensure the buffers are empty */ + kbc_cleanup_buffers(); + + /* keyboard interface test */ + outb(KBC_CMD_KBD_TEST, KBD_COMMAND); + + if (!kbc_output_buffer_full()) { + printk(BIOS_ERR, "Keyboard Interface test timed out.\n"); + return 0; + } + + /* read test result, 0x00 should be returned in case of no failures */ + self_test = inb(KBD_DATA); + + if (self_test != 0x00) { + printk(BIOS_ERR, "Keyboard Interface test failed: 0x%x\n", + self_test); + return 0; + } + return 1; } @@ -147,6 +168,14 @@ static u8 send_keyboard(u8 command) do { if (!kbc_input_buffer_empty()) return 0; outb(command, KBD_DATA); + /* the reset command takes much longer then normal commands and + * even worse, some keyboards do send the ACK _after_ doing the + * reset */ + if (command == 0xFF) { + u8 retries; + for (retries = 9; retries && !kbc_output_buffer_full(); retries--) + ; + } if (!kbc_output_buffer_full()) { printk(BIOS_ERR, "Could not send keyboard command %02x\n", command); @@ -161,6 +190,7 @@ static u8 send_keyboard(u8 command) void pc_keyboard_init(struct pc_keyboard *keyboard) { + u8 retries; u8 regval; if (!CONFIG_DRIVERS_PS2_KEYBOARD) return; @@ -192,10 +222,14 @@ void pc_keyboard_init(struct pc_keyboard } if (regval != KBD_REPLY_ACK) { - printk(BIOS_ERR, "Keyboard selftest failed ACK: 0x%x\n", regval); + printk(BIOS_ERR, "Keyboard reset failed ACK: 0x%x\n", regval); return; } + /* the reset command takes some time, so wait a little longer */ + for (retries = 9; retries && !kbc_output_buffer_full(); retries--) + ; + if (!kbc_output_buffer_full()) { printk(BIOS_ERR, "Timeout waiting for keyboard after reset.\n"); return; @@ -203,7 +237,7 @@ void pc_keyboard_init(struct pc_keyboard regval = inb(KBD_DATA); if (regval != 0xAA) { - printk(BIOS_ERR, "Keyboard selftest failed: 0x%x\n", regval); + printk(BIOS_ERR, "Keyboard reset selftest failed: 0x%x\n", regval); return; } @@ -232,20 +266,20 @@ void pc_keyboard_init(struct pc_keyboard return; } - /* enable the keyboard */ - regval = send_keyboard(0xF4); - if (regval != KBD_REPLY_ACK) { - printk(BIOS_ERR, "Keyboard enable failed ACK: 0x%x\n", regval); - return; - } - /* All is well - enable keyboard interface */ if (!kbc_input_buffer_empty()) return; outb(0x60, KBD_COMMAND); if (!kbc_input_buffer_empty()) return; - outb(0x61, KBD_DATA); /* send cmd: enable keyboard and IRQ 1 */ + outb(0x65, KBD_DATA); /* send cmd: enable keyboard and IRQ 1 */ if (!kbc_input_buffer_empty()) { - printk(BIOS_ERR, "Timeout during final keyboard enable\n"); + printk(BIOS_ERR, "Timeout during keyboard enable\n"); + return; + } + + /* enable the keyboard */ + regval = send_keyboard(0xF4); + if (regval != KBD_REPLY_ACK) { + printk(BIOS_ERR, "Keyboard enable failed ACK: 0x%x\n", regval); return; } }

9 years, 10 months

New patch to review for coreboot: b4145dc Add RS780 defaut graphics ID to AMD Mahogany mainboard.

by Marc Jones

Marc Jones (marcj303(a)gmail.com) just uploaded a new patch set to gerrit, which you can find at http://review.coreboot.org/488 -gerrit commit b4145dcaed06d31ba8c85270e25b810448cd0ea1 Author: Marc Jones <marcj303(a)gmail.com> Date: Wed Dec 14 15:59:02 2011 -0700 Add RS780 defaut graphics ID to AMD Mahogany mainboard. Added the default ID to the mainboard Kconfig. Change-Id: Ie5d39ccdda9d4f5a86214b5bd9ca629070ff152a Signed-off-by: Marc Jones <marcj303(a)gmail.com> --- src/mainboard/amd/mahogany/Kconfig | 4 ++++ src/mainboard/amd/mahogany_fam10/Kconfig | 4 ++++ 2 files changed, 8 insertions(+), 0 deletions(-) diff --git a/src/mainboard/amd/mahogany/Kconfig b/src/mainboard/amd/mahogany/Kconfig index 731002d..b0a46a4 100644 --- a/src/mainboard/amd/mahogany/Kconfig +++ b/src/mainboard/amd/mahogany/Kconfig @@ -73,4 +73,8 @@ config IRQ_SLOT_COUNT int default 11 +config VGA_BIOS_ID + string + default "1002,9615" + endif # BOARD_AMD_MAHOGANY diff --git a/src/mainboard/amd/mahogany_fam10/Kconfig b/src/mainboard/amd/mahogany_fam10/Kconfig index 9c2e081..b8d46eb 100644 --- a/src/mainboard/amd/mahogany_fam10/Kconfig +++ b/src/mainboard/amd/mahogany_fam10/Kconfig @@ -83,4 +83,8 @@ config RAMBASE hex default 0x200000 +config VGA_BIOS_ID + string + default "1002,9615" + endif # BOARD_AMD_MAHOGANY_FAM10

9 years, 10 months

New patch to review for coreboot: 482f534 Fix Fam10 MMCONF_SUPPORT_DEFAULT setting.

by Marc Jones

Marc Jones (marcj303(a)gmail.com) just uploaded a new patch set to gerrit, which you can find at http://review.coreboot.org/487 -gerrit commit 482f5346ca1a49602cf0b3a3f69a22442a25dab1 Author: Marc Jones <marcj303(a)gmail.com> Date: Wed Dec 14 15:33:33 2011 -0700 Fix Fam10 MMCONF_SUPPORT_DEFAULT setting. I misunderstood how kconfig select works. It needs to be selected with a config option. Moved the select to the correct location. Change-Id: If9b1e21e6cbc5af4671efb76cf87dd18dbbe2234 Signed-off-by: Marc Jones <marcj303(a)gmail.com> --- src/cpu/amd/model_10xxx/Kconfig | 3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/src/cpu/amd/model_10xxx/Kconfig b/src/cpu/amd/model_10xxx/Kconfig index 310f98d..221d044 100644 --- a/src/cpu/amd/model_10xxx/Kconfig +++ b/src/cpu/amd/model_10xxx/Kconfig @@ -2,6 +2,7 @@ config CPU_AMD_MODEL_10XXX bool select SSE select SSE2 + select MMCONF_SUPPORT_DEFAULT if CPU_AMD_MODEL_10XXX config CPU_ADDR_BITS @@ -82,6 +83,4 @@ config UPDATE_CPU_MICROCODE patches, so you may need to also disable microcode updates in your operating system in order for this option to matter. -select MMCONF_SUPPORT_DEFAULT - endif # CPU_AMD_MODEL_10XXX

9 years, 10 months

Patch merged into coreboot/master: c18ecfd Lenovo X60/T60: add first_battery setting

by gerrit＠coreboot.org

the following patch was just integrated into master: commit c18ecfd8b74b1a38bee90270e36630f9fad74b98 Author: Sven Schnelle <svens(a)stackframe.org> Date: Wed Dec 7 14:30:58 2011 -0800 Lenovo X60/T60: add first_battery setting The EC allows to select the order in which batteries are (dis)charged. Make this setting available to the user. Change-Id: Id2a98192565419dbb53f3a7cf0b2c46b672a3ed8 Signed-off-by: Sven Schnelle <svens(a)stackframe.org> Build-Tested: build bot (Jenkins) at Wed Dec 7 23:53:28 2011, giving +1 Reviewed-By: Stefan Reinauer <stefan.reinauer(a)coreboot.org> at Fri Dec 9 21:47:10 2011, giving +2 Reviewed-By: Peter Stuge <peter(a)stuge.se> at Wed Dec 14 09:41:46 2011, giving +2 See http://review.coreboot.org/475 for details. -gerrit

9 years, 10 months

Patch merged into coreboot/master: 6cd0925 asus k8v-x: explicitly set RAM and bus voltages

by gerrit＠coreboot.org

the following patch was just integrated into master: commit 6cd0925c8ba4a894ca053fa40477d79122cd6fba Author: Florian Zumbiehl <florz(a)florz.de> Date: Tue Dec 6 10:31:10 2011 +0100 asus k8v-x: explicitly set RAM and bus voltages Change-Id: I9426cafc252ee765d723af569c4a90e090d313d9 Signed-off-by: Florian Zumbiehl <florz(a)florz.de> Build-Tested: build bot (Jenkins) at Sat Dec 10 20:22:34 2011, giving +1 Reviewed-By: Peter Stuge <peter(a)stuge.se> at Wed Dec 14 08:51:58 2011, giving +2 See http://review.coreboot.org/482 for details. -gerrit

9 years, 10 months

Patch merged into coreboot/master: 242cb03 k8: add CONFIG_K8_FORCE_2T_DRAM_TIMING and enable it for asus k8v-x

by gerrit＠coreboot.org

the following patch was just integrated into master: commit 242cb03ce02bde431666e5e6ac7d0eb1375eeba0 Author: Florian Zumbiehl <florz(a)florz.de> Date: Sat Dec 10 19:39:49 2011 +0100 k8: add CONFIG_K8_FORCE_2T_DRAM_TIMING and enable it for asus k8v-x Change-Id: Ia457f92f6fb7e287defb838db07f12d0f1766757 Signed-off-by: Florian Zumbiehl <florz(a)florz.de> Build-Tested: build bot (Jenkins) at Sat Dec 10 20:09:46 2011, giving +1 Reviewed-By: Peter Stuge <peter(a)stuge.se> at Wed Dec 14 08:50:45 2011, giving +2 See http://review.coreboot.org/481 for details. -gerrit

9 years, 10 months

New patch to review for coreboot: cc2e8ec kbd: wait longer for self-test on keyboard reset

by Peter Stuge

Peter Stuge (peter(a)stuge.se) just uploaded a new patch set to gerrit, which you can find at http://review.coreboot.org/486 -gerrit commit cc2e8ec8e1da449915789fe13cc645f23b94669c Author: Mathias Krause <mathias.krause(a)secunet.com> Date: Wed Dec 14 08:40:48 2011 +0100 kbd: wait longer for self-test on keyboard reset Some keyboards take pretty long to respond to a reset command, some even delay the ACK to the command. To make the keyboard driver more robust, increase the timeout for this special command. Also do an interface test after the self-test to ensure the keyboard is functioning properly. Another point is to reenable the keyboard *after* the scancode was set, not before. We also set the system bit when enabling the keyboard because this seems to be what older operating systems do expect. One of the problematic keyboards, which will work with this patch applied, is the DELL RT7D20. Without the patch an overly optimistic operating system, read Linux 2.4, will not recognise the keyboard because coreboot didn't fully initialize it. Change-Id: I28c8e05bdde61f71b7de084c96bc2447c1b9575e Signed-off-by: Mathias Krause <mathias.krause(a)secunet.com> --- src/pc80/keyboard.c | 56 +++++++++++++++++++++++++++++++++++++++++---------- 1 files changed, 45 insertions(+), 11 deletions(-) diff --git a/src/pc80/keyboard.c b/src/pc80/keyboard.c index 8552860..e0a6643 100644 --- a/src/pc80/keyboard.c +++ b/src/pc80/keyboard.c @@ -36,6 +36,7 @@ #define KBC_CMD_READ_COMMAND 0x20 // Read command byte #define KBC_CMD_WRITE_COMMAND 0x60 // Write command byte #define KBC_CMD_SELF_TEST 0xAA // Controller self-test +#define KBC_CMD_KBD_TEST 0xAB // Keyboard Interface test /* The Keyboard controller command byte * BIT | Description @@ -136,6 +137,26 @@ static int kbc_self_test(void) return 0; } + /* ensure the buffers are empty */ + kbc_cleanup_buffers(); + + /* keyboard interface test */ + outb(KBC_CMD_KBD_TEST, KBD_COMMAND); + + if (!kbc_output_buffer_full()) { + printk(BIOS_ERR, "Keyboard Interface test timed out.\n"); + return 0; + } + + /* read test result, 0x00 should be returned in case of no failures */ + self_test = inb(KBD_DATA); + + if (self_test != 0x00) { + printk(BIOS_ERR, "Keyboard Interface test failed: 0x%x\n", + self_test); + return 0; + } + return 1; } @@ -147,6 +168,14 @@ static u8 send_keyboard(u8 command) do { if (!kbc_input_buffer_empty()) return 0; outb(command, KBD_DATA); + /* the reset command takes much longer then normal commands and + * even worse, some keyboards do send the ACK _after_ doing the + * reset */ + if (command == 0xFF) { + u8 retries; + for (retries = 9; retries && !kbc_output_buffer_full(); retries--) + ; + } if (!kbc_output_buffer_full()) { printk(BIOS_ERR, "Could not send keyboard command %02x\n", command); @@ -161,6 +190,7 @@ static u8 send_keyboard(u8 command) void pc_keyboard_init(struct pc_keyboard *keyboard) { + u8 retries; u8 regval; if (!CONFIG_DRIVERS_PS2_KEYBOARD) return; @@ -192,10 +222,14 @@ void pc_keyboard_init(struct pc_keyboard *keyboard) } if (regval != KBD_REPLY_ACK) { - printk(BIOS_ERR, "Keyboard selftest failed ACK: 0x%x\n", regval); + printk(BIOS_ERR, "Keyboard reset failed ACK: 0x%x\n", regval); return; } + /* the reset command takes some time, so wait a little longer */ + for (retries = 9; retries && !kbc_output_buffer_full(); retries--) + ; + if (!kbc_output_buffer_full()) { printk(BIOS_ERR, "Timeout waiting for keyboard after reset.\n"); return; @@ -203,7 +237,7 @@ void pc_keyboard_init(struct pc_keyboard *keyboard) regval = inb(KBD_DATA); if (regval != 0xAA) { - printk(BIOS_ERR, "Keyboard selftest failed: 0x%x\n", regval); + printk(BIOS_ERR, "Keyboard reset selftest failed: 0x%x\n", regval); return; } @@ -232,20 +266,20 @@ void pc_keyboard_init(struct pc_keyboard *keyboard) return; } - /* enable the keyboard */ - regval = send_keyboard(0xF4); - if (regval != KBD_REPLY_ACK) { - printk(BIOS_ERR, "Keyboard enable failed ACK: 0x%x\n", regval); - return; - } - /* All is well - enable keyboard interface */ if (!kbc_input_buffer_empty()) return; outb(0x60, KBD_COMMAND); if (!kbc_input_buffer_empty()) return; - outb(0x61, KBD_DATA); /* send cmd: enable keyboard and IRQ 1 */ + outb(0x65, KBD_DATA); /* send cmd: enable keyboard and IRQ 1 */ if (!kbc_input_buffer_empty()) { - printk(BIOS_ERR, "Timeout during final keyboard enable\n"); + printk(BIOS_ERR, "Timeout during keyboard enable\n"); + return; + } + + /* enable the keyboard */ + regval = send_keyboard(0xF4); + if (regval != KBD_REPLY_ACK) { + printk(BIOS_ERR, "Keyboard enable failed ACK: 0x%x\n", regval); return; } }

9 years, 10 months

Patch set updated for coreboot: e5c75f4 .gitignore util/crossgcc/build-* and unpacked source directories

by Peter Stuge

Peter Stuge (peter(a)stuge.se) just uploaded a new patch set to gerrit, which you can find at http://review.coreboot.org/484 -gerrit commit e5c75f4c5fcd42f4d01deb2a9f4045df1d51e078 Author: Peter Stuge <peter(a)stuge.se> Date: Wed Dec 14 07:32:15 2011 +0100 .gitignore util/crossgcc/build-* and unpacked source directories Change-Id: I85b9dffbbe0c7f1ae8cc2b584196775ba2f816df Signed-off-by: Peter Stuge <peter(a)stuge.se> --- .gitignore | 21 +++++++++++++++++++++ 1 files changed, 21 insertions(+), 0 deletions(-) diff --git a/.gitignore b/.gitignore index 6ea623e..a4bbacd 100644 --- a/.gitignore +++ b/.gitignore @@ -7,6 +7,27 @@ payloads/external/FILO/filo/ payloads/external/SeaBIOS/seabios/ util/crossgcc/tarballs/ util/crossgcc/xgcc/ +util/crossgcc/Python-*/ +util/crossgcc/acpica-unix-*/ +util/crossgcc/binutils-*/ +util/crossgcc/build-binutils/ +util/crossgcc/build-gcc/ +util/crossgcc/build-gdb/ +util/crossgcc/build-gmp/ +util/crossgcc/build-expat/ +util/crossgcc/build-libelf/ +util/crossgcc/build-mpc/ +util/crossgcc/build-mpfr/ +util/crossgcc/build-python/ +util/crossgcc/expat-*/ +util/crossgcc/gcc-*/ +util/crossgcc/gdb-*/ +util/crossgcc/gmp-*/ +util/crossgcc/libelf-*/ +util/crossgcc/mingwrt-*/ +util/crossgcc/mpc-*/ +util/crossgcc/mpfr-*/ +util/crossgcc/w32api-*/ *.o *.elf *.debug

9 years, 10 months

Patch set updated for coreboot: ac891b4 .gitignore util/crossgcc build-* and unpacked source directories

by Peter Stuge

Peter Stuge (peter(a)stuge.se) just uploaded a new patch set to gerrit, which you can find at http://review.coreboot.org/484 -gerrit commit ac891b40fbf5473c14e437795c56871d180d23fa Author: Peter Stuge <peter(a)stuge.se> Date: Wed Dec 14 07:32:15 2011 +0100 .gitignore util/crossgcc build-* and unpacked source directories Change-Id: I85b9dffbbe0c7f1ae8cc2b584196775ba2f816df Signed-off-by: Peter Stuge <peter(a)stuge.se> --- .gitignore | 16 ++++++++++++++++ 1 files changed, 16 insertions(+), 0 deletions(-) diff --git a/.gitignore b/.gitignore index 6ea623e..738f1e6 100644 --- a/.gitignore +++ b/.gitignore @@ -7,6 +7,22 @@ payloads/external/FILO/filo/ payloads/external/SeaBIOS/seabios/ util/crossgcc/tarballs/ util/crossgcc/xgcc/ +util/crossgcc/acpica-unix-*/ +util/crossgcc/binutils-*/ +util/crossgcc/build-binutils/ +util/crossgcc/build-gcc/ +util/crossgcc/build-gdb/ +util/crossgcc/build-gmp/ +util/crossgcc/build-expat/ +util/crossgcc/build-libelf/ +util/crossgcc/build-mpc/ +util/crossgcc/build-mpfr/ +util/crossgcc/build-python/ +util/crossgcc/gcc-*/ +util/crossgcc/gmp-*/ +util/crossgcc/libelf-*/ +util/crossgcc/mpc-*/ +util/crossgcc/mpfr-*/ *.o *.elf *.debug

9 years, 10 months

New patch to review for coreboot: 5858275 .gitignore ectool, inteltool, msrtool, nvramtool and superiotool

by Peter Stuge

Peter Stuge (peter(a)stuge.se) just uploaded a new patch set to gerrit, which you can find at http://review.coreboot.org/485 -gerrit commit 5858275c9f1bee5ee48cea5d614dcba15113e445 Author: Peter Stuge <peter(a)stuge.se> Date: Wed Dec 14 07:39:57 2011 +0100 .gitignore ectool, inteltool, msrtool, nvramtool and superiotool Change-Id: I06e69d97ef3646f79104ec316ce932cc53894c92 Signed-off-by: Peter Stuge <peter(a)stuge.se> --- .gitignore | 9 +++++++++ 1 files changed, 9 insertions(+), 0 deletions(-) diff --git a/.gitignore b/.gitignore index 6ea623e..18ef6d6 100644 --- a/.gitignore +++ b/.gitignore @@ -14,3 +14,12 @@ util/crossgcc/xgcc/ *.\# *.swp *.bin +util/ectool/ectool +util/inteltool/.dependencies +util/inteltool/inteltool +util/msrtool/Makefile +util/msrtool/Makefile.deps +util/msrtool/msrtool +util/nvramtool/.dependencies +util/nvramtool/nvramtool +util/superiotool/superiotool

9 years, 10 months

← Newer
1
...
8
9
10
11
12
13
14
...
18
Older →

Jump to page:

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

coreboot December 2011