|
From: | Richard Henderson |
Subject: | Re: [PATCH v6 00/15] target/riscv: Rationalize XLEN and operand length |
Date: | Mon, 25 Oct 2021 07:58:30 -0700 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.13.0 |
On 10/25/21 2:24 AM, LIU Zhiwei wrote:
I think extending on write is the correct thing.Thanks. I will pick this way.Jumps, exception entry and return, gdb write.If we carefully process jumps and gdb write, I think we can omit exception entry and return. Is it right?
No, though you'd probably have to create a special test case to see it.On exception return, a 64-bit OS can write a 64-bit value into SEPC. But when SEPC is copied to PC during SRET, UXL should be examined and the assignment should extend for a 32-bit user program.
Exception entry would seem to be limited on its face by SXLEN; STVEC will always contain the same number of bits as PC, so no (additional) extension would be required. But we do have to be careful to interpret the target_ulong value properly for the current SXLEN.
r~
[Prev in Thread] | Current Thread | [Next in Thread] |