summaryrefslogtreecommitdiff
path: root/arch
Commit message (Collapse)AuthorAgeFilesLines
* FogBugz #385884: Configuring SDRAM extra cycles on timing parametersrel_socfpga_v2013.01.01_17.06.02_prrel_socfpga_v2013.01.01_17.06.01_prrel_socfpga_v2013.01.01_17.05.02_prrel_socfpga_v2013.01.01_17.05.01_prrel_socfpga_v2013.01.01_17.04.02_prrel_socfpga_v2013.01.01_17.04.01_prrel_socfpga_v2013.01.01_17.03.01_prrel_socfpga_v2013.01.01_17.02.01_prrel_socfpga_v2013.01.01_16.10.02_prrel_socfpga_v2013.01.01_16.10.01_prACDS17.0_REL_GSRD_RC1ACDS17.0_REL_GSRD_PRACDS16.1_REL_GSRD_RC7ACDS16.1_REL_GSRD_RC6ACDS16.1_REL_GSRD_RC5ACDS16.1_REL_GSRD_RC4ACDS16.1_REL_GSRD_PRChin Liang See2016-09-262-1/+42
| | | | | | | | | | | To enable configuration of sdr.ctrlcfg.extratime1 register which enable extra clocks for read to write command timing. This is critical to ensure successful LPDDR2 interface Signed-off-by: Chin Liang See <clsee@altera.com> --- Changes in v2 - Remove the change of handoff as our dev kit using DDR3
* FogBugz #353914: Fix SPL FPGA configuration with bridge enabledrel_socfpga_v2013.01.01_16.08.02_prChin Liang See2016-07-291-0/+1
| | | | | | | | | To fix the Preloader to ensure the bridges are disabled when programming the FPGA. The bridge is now enabled upon Reset Manager initialization but left open even during FPGA programming which is not safe to do so. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #353914: Fix SDRAM self refresh failurerel_socfpga_v2013.01.01_16.08.01_prrel_socfpga_v2013.01.01_16.07.02_prrel_socfpga_v2013.01.01_16.07.01_prrel_socfpga_v2013.01.01_16.06.02_prrel_socfpga_v2013.01.01_16.06.01_prrel_socfpga_v2013.01.01_16.05.02_prrel_socfpga_v2013.01.01_16.05.01_prrel_socfpga_v2013.01.01_16.04.02_prrel_socfpga_v2013.01.01_16.04.01_prACDS16.0_REL_GSRD_RC7ACDS16.0_REL_GSRD_RC6ACDS16.0_REL_GSRD_RC5ACDS16.0_REL_GSRD_RC4ACDS16.0_REL_GSRD_PRChin Liang See2016-03-241-2/+7
| | | | | | | | | | | | To fix the broken SDRAM self refresh feature where the SDRAM content is wiped off after warm reset. The SDRAM ECC initialization should be skipped when SDRAM self refresh is enabled and after a warm reset. Signed-off-by: Chin Liang See <clsee@altera.com> --- Changes for v2 - Update the comment to reflect the code change
* FogBugz #359965: Enforce full SDRAM ECC init when ECC is enabledrel_socfpga_v2013.01.01_16.03.03_prACDS16.0_REL_GSRD_RC3ACDS16.0_REL_GSRD_RC2ACDS16.0_REL_GSRD_RC1Chin Liang See2016-03-073-180/+14
| | | | | | | | | | | To enforce initialization of entire SDRAM ECC bit if SDRAM ECC is enabled. This is to avoid any HW complication and build error if incorrect combinations are selected in tools. Signed-off-by: Chin Liang See <clsee@altera.com> --- Changes for v2 - Fix spelling error in printout
* arm: Switch to -mno-unaligned-access when supported by the compilerrel_socfpga_v2013.01.01_16.03.02_prrel_socfpga_v2013.01.01_16.03.01_prrel_socfpga_v2013.01.01_16.02.02_prrel_socfpga_v2013.01.01_16.02.01_prrel_socfpga_v2013.01.01_16.01.02_prrel_socfpga_v2013.01.01_16.01.01_prChin Liang See2015-12-281-0/+1
| | | | | | | | | | | | | When we tell the compiler to optimize for ARMv7 (and ARMv6 for that matter) it assumes a default of SCTRL.A being cleared and unaligned accesses being allowed and fast at the hardware level. We set this bit and must pass along -mno-unaligned-access so that the compiler will still breakdown accesses and not trigger a data abort. Cc: Albert ARIBAUD <albert.u.boot@aribaud.net> Cc: Mans Rullgard <mans@mansr.com> Signed-off-by: Tom Rini <trini@ti.com> Signed-off-by: Chin Liang See <clsee@altera.com>
* common: board_f: cosmetic use __weak for ledsChin Liang See2015-12-091-19/+9
| | | | | | | | | | | First of all this looks a lot better, but it also prevents a gcc warning (W=1), that the weak function has no previous prototype. cc: Simon Glass <sjg@chromium.org> Signed-off-by: Jeroen Hofstee <jeroen@myspectrum.nl> Acked-by: Simon Glass <sjg@chromium.org> Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #297183: Enabling ECC overwrite when SDRAM ECC is enabledrel_socfpga_v2013.01.01_15.11.02_prrel_socfpga_v2013.01.01_15.11.01_prrel_socfpga_v2013.01.01_15.10.02_prrel_socfpga_v2013.01.01_15.10.01_prrel_socfpga_v2013.01.01_15.09.02_prACDS15.1.1_REL_GSRD_RC6ACDS15.1.1_REL_GSRD_RC5ACDS15.1.1_REL_GSRD_RC4ACDS15.1.1_REL_GSRD_RC3ACDS15.1.1_REL_GSRD_RC2ACDS15.1.1_REL_GSRD_RC10ACDS15.1.1_REL_GSRD_RC1ACDS15.1.1_REL_GSRD_PRChin Liang See2015-08-261-0/+6
| | | | | | | | Enable SDRAM ECC overwrite where it occurs when a correctable ECC error is seen. A new read/modify/write to be scheduled for that location to clear the ECC error. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #302866: Enabling ECC initialization support for 2GB SDRAMChin Liang See2015-08-261-6/+6
| | | | | | | | | | | | | Increase the buffer size to store the DMA microcode. These microcode are used to initialize the SDRAM ECC bit. The size of microcode will grow bigger with size of SDRAM. Signed-off-by: Chin Liang See <clsee@altera.com> --- Changes for v2 - Fix some grammer in commit message - Renew the copyright header - Increase the buffer size for storing the microcode
* FogBugz #260361: Fixed intermittent QSPI flash erase failuresTien Fong Chee2015-05-081-2/+2
| | | | | | | | | This patch fixed the overflow formula in both get_timer_masked and get_timer_count_masked. This bug can cause QSPI flash failed to erase and time out intermittent during the process of erasing. Signed-off-by: Tien Fong <tfchee@altera.com>
* FogBugz #231136: Skip pll workaround if booting from fpgarel_socfpga_v2013.01.01_15.05.03_prrel_socfpga_v2013.01.01_15.05.02_prrel_socfpga_v2013.01.01_15.05.01_prrel_socfpga_v2013.01.01_15.04.02_prrel_socfpga_v2013.01.01_15.04.01_prrel_socfpga_v2013.01.01_15.03.02_prrel_socfpga_v2013.01.01_15.03.01_prrel_socfpga_v2013.01.01_15.02.02_prrel_socfpga_v2013.01.01_15.02.01_prrel_socfpga_v2013.01.01_15.01.02_prrel_socfpga_v2013.01.01_15.01.01_prrel_socfpga_v2013.01.01_14.12.01_prrel_socfpga_v2013.01.01_14.11.02_prACDS15.0_REL_GSRD_PRVince Bridgers2014-11-122-1/+16
| | | | | | | | | This patch skips the PLL workaround for ramboot if booting from FPGA, and adds the preloader generator option CONFIG_PRELOADER_RAMBOOT_PLLRESET to force the workaround from the preloader if requested by the user, or configured by editing the configuration. Signed-off-by: Vince Bridgers <vbridger@opensource.altera.com>
* ARM: Add workaround for Cortex-A9 errata 761320Nitin Garg2014-11-051-0/+5
| | | | | | | | | Full cache line writes to the same memory region from at least two processors might deadlock the processor. Exists on r1, r2, r3 revisions. Signed-off-by: Nitin Garg <nitin.garg@freescale.com> Acked-by: Fabio Estevam <fabio.estevam@freescale.com>
* ARM: Add workaround for Cortex-A9 errata 794072Nitin Garg2014-11-051-1/+1
| | | | | | | | | A short loop including a DMB instruction might cause a denial of service on another processor which executes a CP15 broadcast operation. Exists on r1, r2, r3, r4 revisions. Signed-off-by: Nitin Garg <nitin.garg@freescale.com> Acked-by: Dirk Behme <dirk.behme@de.bosch.com>
* ARM: implement erratum 716044 workaroundStephen Warren2014-11-051-0/+25
| | | | | | | Add common code to enable the workaround for ARM erratum 716044. This will be enabled for Tegra. Signed-off-by: Stephen Warren <swarren@nvidia.com>
* FogBugz #241780: Disable ECC, add preprocessor checks for consistent staterel_socfpga_v2013.01.01_14.11.01_prVince Bridgers2014-10-311-9/+25
| | | | | | | | | | | | | | | | Somehow, ECC was enabled but scrubbing was disabled for our git repo default board configuration. This patch corrects that issue by setting CONFIG_HPS_SDR_CTRLCFG_CTRLCFG_ECCEN to 0 for the default configuration, makes sure CONFIG_HPS_SDR_CTRLCFG_STATICCFG_USEECCASDATA is always disabled, and CONFIG_HPS_SDR_CTRLCFG_CTRLCFG_ECCCORREN is also always disabled. Note that the SDRAM configuration code always sets CONFIG_HPS_SDR_CTRLCFG_STATICCFG_USEECCASDATA and CONFIG_HPS_SDR_CTRLCFG_CTRLCFG_ECCCORREN to 0 overriding the preloader generator handoffs leaving an OS specific driver with the choice of using those features if needed. The preloader does the basic SDRAM ECC initialization tasks required for basic ECC initialization and protection. Signed-off-by: Vince Bridgers <vbridger@opensource.altera.com>
* FogBugz #227886: Setup SDRAM Protection Rule to prevent access beyond max memrel_socfpga_v2013.01.01_14.10.02_prACDS14.1_REL_GSRD_PR_UPDATE2ACDS14.1_REL_GSRD_PR_UPDATE1ACDS14.1_REL_GSRD_PRACDS14.0.2_REL_SGMII_PRACDS14.0.2_REL_GSRD_PRVince Bridgers2014-10-084-4/+411
| | | | | | | | | | | | | This patch enhances the previously submitted patch for FogBugz #223741 by programming the SDRAM protection rules to prevent SDRAM accesses beyond maximum configured SDRAM. This patch also corrects an issue in the way SDRAM rows was calculated and configured in that same patch. The number of rows is set such that 4G of memory is decoded to workaround the SDRAM errata and limits memory accesses by using the SDRAM range protection logic to cause a data abort if an access occurs beyond available, configured memory in the system. Signed-off-by: Vince Bridgers <vbridger@opensource.altera.com>
* Fogbugz #223877: Force globals into .data segment to compact memory footprintrel_socfpga_v2013.01.01_14.10.01_prrel_socfpga_v2013.01.01_14.09.02Vince Bridgers2014-09-152-6/+6
| | | | | | | | | | Use linker directives to force certain global variables into .data segment instead of bss, so bss section is not used prior to u-boot relocation.. Signed-off-by: Vince Bridgers <vbridger@opensource.altera.com> --- V3: add more information to commit log V2: Change emails to opensource emails
* FogBugz #230385: Replace CONFIG_PRELOADER_WARMRST_PRESERVE_SDRAM with ↵Chin Keong Ang2014-09-142-3/+3
| | | | | | | | | | | | | | CONFIG_HPS_RESET_WARMRST_HANDSHAKE_SDRAM To enable SDRAM self refresh support, CONFIG_PRELOADER_WARMRST_PRESERVE_SDRAM and CONFIG_HPS_RESET_WARMRST_HANDSHAKE_SDRAM should always be set to exactly the same value. Replacing CONFIG_PRELOADER_WARMRST_PRESERVE_SDRAM with CONFIG_HPS_RESET_WARMRST_HANDSHAKE_SDRAM avoided the posibility of making a mistake. Signed-off-by: Chin Keong Ang <ckang@altera.com>
* Fogbugz #229396 Fix preloader handling of BSEL 0rel_socfpga_v2013.01.01_14.09.01ACDS14.0.1_REL_GSRD_RC2Matthew Gerlach2014-09-102-0/+17
| | | | | | | | | | | | When BootSelect is set to zero. The BootRom does almost nothing. The BootRom does not setup any of the handoff to the preloader, nor does the BootROM perform any work necessary to handle a warm reset. This patch fixes the preloader's handling of BootSelect being zero. When BootSelect is zero, the preloader will not examine any handoff from the BootRom, and the preloader will execute a full cold reset startup sequence. Signed-off-by: Matthew Gerlach <mgerlach@opensource.altera.com>
* Fogbugz #197088: Configure Debug Clock source from Quartus/Qsys HandoffVince Bridgers2014-09-032-0/+11
| | | | | | | | | This patch configures the debug clock source according to the handoff from Quartus. Signed-off-by: Vince Bridgers <vbridger@opensource.altera.com> --- V2: Change email references to opensource emails
* Fogbugz #223741: Workaround SDRAM controller hang when reading beyond max memVince Bridgers2014-09-031-9/+74
| | | | | | | | | | | | | | | | | This patch prevents the SDRAM controller from hanging indeterminately if a memory read operation is posted to the SDRAM controller beyond the amount of memory controlled by the SDRAM controller. Signed-off-by: Vince Bridgers <vbridger@opensource.altera.com> --- V5: - Read sdram register and change values back to handoff values before calculating sdram size - Add informative message to user about sdram addrorder change V4: - Remove extra linefeeds, remove references to FBs in source code, restate commit header again. V3: - Change emails to opensource emails V2: - Updates per first Crucible review - Simplify commit log entry
* FogBugz #194224: Enable SDRAM self refresh supportChin Keong Ang2014-08-219-95/+320
| | | | | | | | | | To enable SDRAM contents preservation over warm reset. Enable by defining CONFIG_PRELOADER_WARMRST_PRESERVE_SDRAM as 1 in build.h Once enabled, during warm reset, it will skip resetting SDRAM controller and skip calibrating SDRAM PHY. Signed-off-by: Chin Keong Ang <ckang@altera.com>
* FogBugz #213213: self refresh settingsrel_socfpga_v2013.01.01_14.08.02ACDS14.0.1_REL_GSRD_RC3ACDS14.0.1_REL_GSRD_PRAlan Tull2014-08-192-1/+38
| | | | | | | | | | | | | Settings to configure the sdram controller for when we want to enable ddr self-refresh. sdr.ctrlcfg.lowpwreq.selfrfshmask = 3 sdr.ctrlcfg.lowpwrtiming.clkdisablecycles = 8 sdr.ctrlcfg.dramtiming4.selfrfshexit = 512 Signed-off-by: Alan Tull <atull@altera.com> v2: selfrfshmask = 3 instead of 1 to chose all ddr
* FogBugz #210587: Fixing PLL HW configuration issuerel_socfpga_v2013.01.01_14.08.01rel_socfpga_v2013.01.01_14.07.02rel_socfpga_v2013.01.01_14.07.01rel_socfpga_v2013.01.01_14.06.02Chin Liang See2014-06-171-42/+10
| | | | | | | To update Clock Manager driver to avoid PLL from using external regulator.At same time, removed double reset code workaround. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #208554: RAM boot to reset Clock Manager during CSEL 0rel_socfpga_v2013.01.01_14.06.01Chin Liang See2014-06-036-2/+130
| | | | | | | | | | | | | | | To enable RAM boot when CSEL = 0. The RAM boot code will put Main PLL and Peripheral to bypass mode. It will set the QSPI and NANDSDMMC clock dividers for both PLLs to default value. Later, the code will disable the RAM boot before triggering warm reset. The RAM boot code is located at the last 4kB of on-chip RAM. The purpose of this RAM boot is to overcome boot up hang issue which due to faulty PLL. Signed-off-by: Chin Liang See <clsee@altera.com> --- Changes for v2 - Auto calculate the function size when copying the function
* FogBugz #208110: Fix the L4 clock calculation algorithmChin Liang See2014-06-021-1/+1
| | | | | | | To fix the L4 clock calculation algorithm as the L4 divider in value of power of 2 instead divisor itself. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #207596: Fix LWH2F bridge issue when boot from FPGAChin Liang See2014-05-293-6/+4
| | | | | | | | | To fix the issue where the LWH2F bridge is not enabled per handoff when boot from FPGA. This will ensure Preloader pass the bridge handoff info to U-Boot correctly.At same time, this patch also fixed the label duplication within lowlevel_init.S. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #192038: Fix SDRAM PLL Phase AligningChin Liang See2014-05-151-3/+39
| | | | | | | | | | | To fix the SDRAM PLL phase aligning issue which cause SDRAM calibration failed. This patch will reset the SDRAM VCO twice and implemented software de-bouncing check for PLL lock status. Signed-off-by: Chin Liang See <clsee@altera.com> --- Changes for v2 - Using udelay instead for loop implementation
* FogBugz #191875: Enable U-Boot command line editingChin Liang See2014-03-241-0/+18
| | | | | | | | Enhance U-Boot command console experience by enabling command line editing. User can use up arrow to retrieve previous typed command and edit it later on. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #188622: Ensure BootROM works if SDMMC PWREN is usedrel_socfpga_v2013.01.01_14.03.02Chin Liang See2014-03-183-0/+21
| | | | | | | | If SDMMC PWREN is used, we need to ensure BootROM always reconfigure IOCSR and pinmux after warm reset. This is to cater the use case of board design which is using SDMMC PWREN pins. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #189084: Enhance U-Boot to configure L2 address filterChin Liang See2014-03-131-7/+6
| | | | | | | Enhance the U-Boot to configure the L2 address filter. This is to avoid the dependency of getting Preloader to configure it. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #121124:Fix DW USB driver failing to run on board.rel_socfpga_v2013.01.01_14.03.01rel_socfpga_v2013.01.01_14.02.02Tien Hock Loh2014-02-171-0/+2
| | | | | | | | | This change fixes PID toggling so that USB transaction works correctly, updating the codes so that it conforms to the u-boot standard better, ie. replacing dwc_write32 with __raw_write. Also replaces infinite loop with time outs. Signed-off-by: Tien Hock Loh <thloh@altera.com>
* FogBugz #182048: Fix sdram_applycfg_ocram functionChin Liang See2014-02-111-2/+3
| | | | | | | | | To ensure the function always return 0 upon completion. This is to avoid the user impression that the function is hitting error due to non zero return. At same time, fix the code to enable back the i-cache and branch predictor after applycfg is applied. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #182047: Fix main_nand_sdmmc_clk clock configurationChin Liang See2014-02-113-1/+5
| | | | | | | | To fix the Clock Manager driver which currently not configuring the main_nand_sdmmc_clk as its not used. This patch will enable user to have both NAND and SDMMC clock from different PLL sources. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #169625: Enhance Preloader programming FPGA from SDMMCChin Liang See2014-02-031-2/+119
| | | | | | | | | | | | | | Enhance the Preloader to have the capability of programming FPGA from SD card's FAT partition. User just need to put the RBF file into FAT partition. This is an additional feature on top of existing Preloader programming FPGA from QSPI flash. Signed-off-by: Chin Liang See <clsee@altera.com> --- Changes for v3 - Used puts for prinout without argument Changes for v2 - Removed debug printf
* FogBugz #158674: Enabling I2C support for U-BootChin Liang See2014-01-223-10/+11
| | | | | | | | | | | | | | | | Enable the DesignWare I2C controller support within U-Boot. Only I2C0 is enabled per Cyclone V and Arria V dev kit. At same time, rename socfpga_base_addrs.h to hardware.h. This is to align with the file name used by DesignWare I2C driver. Signed-off-by: Chin Liang See <clsee@altera.com> --- Changes for v4 - Update the license header for hardware.h Changes for v3 - Renamed socfpga_base_addrs.h to hardware.h Changes for v2 - Updated hardware.h which mainly used for I2C driver
* FogBugz #178339: Clearing PLL loss bits after clock configurationChin Liang See2014-01-202-0/+8
| | | | | | | | Clear all the PLL lost of lock bits in Clock Manager's interrupt status register. The false PLL lost of lock happen during the PLL reconfiguration in Preloader stage. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #158674: Enhance drivers to use derived clock valueChin Liang See2014-01-104-28/+53
| | | | | | | | | | | | | | | | Enhance the drivers for flash controllers (SDMMC and QSPI) and UART to determine the incoming clock frequency (from PLL) based on derived values instead of macros. The derived values are based on Clock Manager registers and incoming clock source frequencies (from handoff file pll_config.h). This enhancement will avoid user manually update the clock frequencies for drivers when any changes made to Clock Manager configuration. Signed-off-by: Chin Liang See <clsee@altera.com> --- Changes for v2 - Used uint32_t instead unsigned long for Clock Manager register struct
* FogBugz #158674: Enhance Preloader to display clock infoChin Liang See2014-01-103-0/+260
| | | | | | | | Preloader will display the clock info such as MPU clock, DDR clock and others during run time. It will help user to know the current clock configuration being used without reading the handoff file. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #159721: Enhance Arria V MPU clock to 1050MHzChin Liang See2014-01-102-0/+14
| | | | | | | | Enhance the Arria V MPU clock from 800MHz to 1050MHz. This is to align with the new Arria V dev kit which is using I3 speed grade parts. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #173924: Adding SDRAM ECC Scrubbing using DMAChin Liang See2014-01-103-0/+193
| | | | | | | | | | | | | | | User can optional enable the Preloader to scrub the SDRAM memory to initialize the SDRAM ECC bits. This is to enable subsequent software that running on SDRAM enable the SDRAM ECC interrupt. To do the scrubbing, DMA is used to speed up the scrubbing speed. At same time, the DMA run in parallel while Preloader loading the subsequent boot image from flashes. To scrub the whole 1GB memory, it took 1446ms. As comparison, scrubbing using processor and memset function is taking 13s. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #173924: Adding DMA PL330 supportChin Liang See2014-01-102-0/+35
| | | | | | | | | | To add the driver support for DMA PL330 that located within HPS. The driver does have high level function which enable memory-to-memory, memory-to-peripheral and peripheral-to-memory transfer. It also include a memory scrubbing function which write zeroes to the assigned memory region Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #174257: Enhance fpga2sdram bridge release sequenceChin Liang See2014-01-022-3/+3
| | | | | | | | | | Enhance the fpga2sdram bridge release sequence to support asymmetric multiprocessing (AMP) use case. The master in FPGA might sending transaction to fpga2sdram bridge before the bridge is enabled. This enhancement will ensure the transaction will be back pressured and continued normally once the bridge is enabled. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #158875: Enhance semihosting code for THUMB2 modeChin Liang See2013-12-191-0/+4
| | | | | | | Enhance the semihosting code where we will using different SVC value for THUMB mode. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #166933: Fixed OCRAM false double bit errorChin Liang See2013-12-111-0/+14
| | | | | | | | | | Ensure the stack pointer memory region is initialized during the early run of the Preloader and U-Boot. This is to avoid the false double bit error which might occurred during stack access. It is due to the OCRAM controller always read in 64 bits manner. This issue is spotted when we switch to new Altera version GNU toolchain. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #168357: Enhance to auto calculate SDRAM sizeChin Liang See2013-12-034-8/+64
| | | | | | | | | | | | | Enhance Preloader and U-Boot to auto calculate the SDRAM size. This will save user from the hassle of modifying the PHYS_SDRAM_1 macro when the SDRAM change. The SDRAM size calculation will based on SDRAM controller settings such as address row, columns, banks and others. Signed-off-by: Chin Liang See <clsee@altera.com> --- Changes for v2 - using macro for SDRAM width with ECC enabled - fixed the SDRAM size calculation algorithm
* FogBugz #161845: Fix build error when SDRAM ECC disabled and FAT enabledChin Liang See2013-12-021-0/+12
| | | | | | | | To fix the build error during the comhination where SDRAM ECC is disabled and Preloader FAT support is enabled. The SDRAM ECC padding is not required when the SDRAM ECC is disabled. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #163901: Display message when watchdog is enabledChin Liang See2013-12-021-0/+3
| | | | | | | | This is to display a message at UART terminal when the watchdog is enabled. It will remind user the possibility of system reset caused by watchdog timeout. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #161193: Enhancing "bridge_enable_handoff" commandChin Liang See2013-11-271-14/+2
| | | | | | | | | | | | | | | Bridge_enable_handoff command is used to enable all the bridges based on handoff values. It took place after the FPGA programming. This enhancement will remove all memory display commands as they will trigger data abort if user press enter many times after run bridge_enable_handoff. The data abort is caused by U-Boot memory display features which will auto increment the address after user press enter later on. When this happen, the memory display command will display next address which are not a valid address and causing data abort. Signed-off-by: Chin Liang See <clsee@altera.com>
* FogBugz #136143: Enhancing Preloader SDRAM memory testChin Liang See2013-11-145-22/+127
| | | | | | | | | | | | | | | | | To enhance the existing SDRAM memory test. Added a memory test which is utilizing the PRBS (pseudo-random binary sequence) algorithm. At same time, ensure lowest memory map is mapped to SDRAM so it can be tested too. Signed-off-by: Chin Liang See <clsee@altera.com> --- Changes for v4 - Removed redundant code Changes for v3 - reshuffled the condition check for coverage within SDRAM test Changes for v2 - used macro to determine the SDRAM test coverage - updated comments
* FogBugz #163055: Fixing Preloader FPGA programmingChin Liang See2013-11-122-13/+16
| | | | | | | | | To fix the Preloader FPGA programming issue due to recent enhancement. This is to avoid programming additional words into FPGA. At same time, update the FPGA2SDRAM port release code after the FPGA programming. Signed-off-by: Chin Liang See <clsee@altera.com>