Libfdt support & local Timer node addition. #112
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
BBL performs in memory operations to modify FDT entries. This is perfectly fine for existing use case where only some properties are modified. However, this is not an feasible option in case of multiple subnode addition. Add libfdt support by linking the libfdt source directly. I understand that libfdt might have increased the size of bbl significantly. However, this is just a temporary patch only meant for linux capable boards. Once we have a proper boot loader with separate firmware layer, this patch will no longer be required. For example, u-boot has already support for libfdt. Thus, this patch will be redundant and can be removed. Signed-off-by: Atish Patra <[email protected]>
The present device tree is loaded from ROM and doesn't have an entry for local timer. Linux kernel needs a local timer node so that timer interrupt can be registered as a regular interrupt under HLIC irq domain. Signed-off-by: Atish Patra <[email protected]>
|
FYI: I have reused code from glibc in util/bcopy.c & string.c. It was all under openBSD license. So I thought it would be okay to directly include. If not, please advise. |
atishp04
added a commit
to atishp04/riscv-linux
that referenced
this pull request
Jul 9, 2018
Currently, ther timer interrupt handleri(riscv_timer_interrupt()) is invoked directly from the arch handler. This approach creates interwinding of driver and architecture code which is not ideal. Register timer interrupt as a per cpu based irq behind generic Linux IRQ infrastructure. As the local timer node is directly connected to individual harts, HLIC acts as the parent irq domain. This patch requires necessary bbl changes that adds timer node in device tree which can be found here. riscv-software-src/riscv-pk#112 Before the patch --------------------------------------------------- CPU1 CPU2 CPU3 CPU4 21: 12 8 6 5 riscv,plic0,c000000 53 eth0 37: 0 0 0 0 riscv,plic0,c000000 32 xilinx-pcie 43: 0 0 0 0 riscv,plic0,c000000 4 10010000.serial 44: 0 0 0 0 riscv,plic0,c000000 5 10011000.serial 45: 0 0 0 0 riscv,plic0,c000000 51 10040000.spi 46: 0 0 0 0 riscv,plic0,c000000 52 10041000.spi 47: 25 1 26 50 riscv,plic0,c000000 6 10050000.spi After the patch --------------------------------------------------- CPU1 CPU2 CPU3 CPU4 5: 271 0 0 0 riscv,cpu_intc,1 5 local_timer 6: 0 307 0 0 riscv,cpu_intc,2 5 local_timer 7: 0 0 303 0 riscv,cpu_intc,3 5 local_timer 8: 0 0 0 223 riscv,cpu_intc,4 5 local_timer 47: 337 489 389 35 riscv,plic0,c000000 4 10010000.serial 48: 0 0 0 0 riscv,plic0,c000000 5 10011000.serial 49: 0 0 0 0 riscv,plic0,c000000 51 10040000.spi 50: 0 0 0 0 riscv,plic0,c000000 52 10041000.spi 51: 2 14 47 39 riscv,plic0,c000000 6 10050000.spi Signed-off-by: Atish Patra <[email protected]>
atishp04
added a commit
to atishp04/riscv-linux
that referenced
this pull request
Jul 9, 2018
Currently, the timer interrupt handler(riscv_timer_interrupt()) is invoked directly from the arch handler. This approach creates interwinding of driver and architecture code which is not ideal. Even if the all timer hwirq have same irq number, they are discrete interrupts connected to each core. Moreover, HLIC is modelled in a way that there are individual HLIC per hart. Thus, register timer interrupt as an individual irq under generic Linux IRQ management framework instead of per-cpu irq. Individual HLIC acts as the parent irq domain. This patch requires necessary bbl changes that adds timer node in device tree which can be found here. riscv-software-src/riscv-pk#112 Before the patch --------------------------------------------------- CPU1 CPU2 CPU3 CPU4 21: 12 8 6 5 riscv,plic0,c000000 53 eth0 37: 0 0 0 0 riscv,plic0,c000000 32 xilinx-pcie 43: 0 0 0 0 riscv,plic0,c000000 4 10010000.serial 44: 0 0 0 0 riscv,plic0,c000000 5 10011000.serial 45: 0 0 0 0 riscv,plic0,c000000 51 10040000.spi 46: 0 0 0 0 riscv,plic0,c000000 52 10041000.spi 47: 25 1 26 50 riscv,plic0,c000000 6 10050000.spi After the patch --------------------------------------------------- CPU1 CPU2 CPU3 CPU4 5: 271 0 0 0 riscv,cpu_intc,1 5 local_timer 6: 0 307 0 0 riscv,cpu_intc,2 5 local_timer 7: 0 0 303 0 riscv,cpu_intc,3 5 local_timer 8: 0 0 0 223 riscv,cpu_intc,4 5 local_timer 47: 337 489 389 35 riscv,plic0,c000000 4 10010000.serial 48: 0 0 0 0 riscv,plic0,c000000 5 10011000.serial 49: 0 0 0 0 riscv,plic0,c000000 51 10040000.spi 50: 0 0 0 0 riscv,plic0,c000000 52 10041000.spi 51: 2 14 47 39 riscv,plic0,c000000 6 10050000.spi Signed-off-by: Atish Patra <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds support for libfdt and adds local timer node into the device tree. It is just a temporary patch only meant for linux capable boards. I hope future replacement of bbl will contain libfdt support inbuilt. Thus 1st patch will not be required afterwards.
That's why it is rebased on the commit id 5a0e3e5 which was referred in freedom-u-sdk.
Please let me know if I should rebase these patches on top of some other commit.