Re: [PATCH v23 06/28] riscv/mm : ensure PROT_WRITE leads to VM_READ | VM_WRITE

[Deepak Gupta]

On Sat, May 30, 2026 at 8:02 AM Andreas Schwab <schwab@xxxxxxxxxxxxxx> wrote:
>
> On Nov 12 2025, Deepak Gupta via B4 Relay wrote:
>
> > From: Deepak Gupta <debug@xxxxxxxxxxxx>
> >
> > `arch_calc_vm_prot_bits` is implemented on risc-v to return VM_READ |
> > VM_WRITE if PROT_WRITE is specified. Similarly `riscv_sys_mmap` is
> > updated to convert all incoming PROT_WRITE to (PROT_WRITE | PROT_READ).
> > This is to make sure that any existing apps using PROT_WRITE still work.
> >
> > Earlier `protection_map[VM_WRITE]` used to pick read-write PTE encodings.
> > Now `protection_map[VM_WRITE]` will always pick PAGE_SHADOWSTACK PTE
> > encodings for shadow stack. Above changes ensure that existing apps
> > continue to work because underneath kernel will be picking
> > `protection_map[VM_WRITE|VM_READ]` PTE encodings.
>
> This breaks LTP mmap04:
>
> $ ./mmap04
> tst_test.c:2042: TINFO: LTP version: 20260529.5ccf816f
> tst_test.c:2045: TINFO: Tested kernel: 7.0.10-5-default #1 SMP PREEMPT_DYNAMIC Sat May 23 12:09:09 UTC 2026 (bb95589) riscv64
> tst_kconfig.c:90: TINFO: Parsing kernel config '/proc/config.gz'
> tst_kconfig.c:753: TINFO: CONFIG_FAULT_INJECTION kernel option detected which might slow the execution
> tst_test.c:1870: TINFO: Overall timeout per run is 0h 04m 00s
> mmap04.c:66: TPASS: mapping permissions in /proc matched: ---p
> mmap04.c:66: TPASS: mapping permissions in /proc matched: ---s
> mmap04.c:66: TPASS: mapping permissions in /proc matched: r--p
> mmap04.c:66: TPASS: mapping permissions in /proc matched: r--s
> mmap04.c:68: TFAIL: mapping permissions in /proc mismatched, expected: -w-p, found: rw-p
> mmap04.c:68: TFAIL: mapping permissions in /proc mismatched, expected: -w-s, found: rw-s
> mmap04.c:66: TPASS: mapping permissions in /proc matched: rw-p
> mmap04.c:66: TPASS: mapping permissions in /proc matched: rw-s
> mmap04.c:66: TPASS: mapping permissions in /proc matched: r-xp
> mmap04.c:66: TPASS: mapping permissions in /proc matched: r-xs
> mmap04.c:68: TFAIL: mapping permissions in /proc mismatched, expected: -wxp, found: rwxp
> mmap04.c:68: TFAIL: mapping permissions in /proc mismatched, expected: -wxs, found: rwxs
> mmap04.c:66: TPASS: mapping permissions in /proc matched: rwxp
> mmap04.c:66: TPASS: mapping permissions in /proc matched: rwxs

But what does a writable only mapping even mean? It can always be read.
I see two ways forward:

- It can be done in generic way where incoming PROT_WRITE means
  PROT_READ | PROT_WRITE irrespective of RISC-V. Although others
  (x86, arm, etc) would have to weigh in.

OR

- mmap04 LTP test can be updated to expect either of "rw-p" or "-w-p"
  whenever only PROT_WRITE was specified.

I think latter (updating mmap04 test) is the easier path.

>
> Summary:
> passed   10
> failed   4
> broken   0
> skipped  0
> warnings 0
>
> --
> Andreas Schwab, schwab@xxxxxxxxxxxxxx
> GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
> "And now for something completely different."