Numbering rows within group in a data frame when multiple variables grouped

Question

Yes, there are several questions similar to this, but none of them involve multiple grouped variables, which causes the solutions for those questions to not work properly. The best similar questions I can find are:

1. Numbering rows within groups in a data frame, and
2. Create a sequential number (counter) for rows within each group of a dataframe [duplicate]

Dummy data for my case:

library(dplyr)
df <- tibble(
    s1 = c("111", "111", "111", "112", "112", "114", "114", "115"),
    s2 = c(rep("A", 5), rep("B", 3)),
    val = rnorm(8)
)

I want to provide a grouping ID for s1 within group s2. That is, I want it to reset each time s2 changes. Desired output:

# A tibble: 8 x 4
  s1    s2        val    id
  <chr> <chr>   <dbl> <dbl>
1 111   A     -0.465      1
2 111   A      0.871      1
3 111   A      0.823      1
4 112   A      0.561      2
5 112   A      0.197      2
6 114   B     -0.743      1
7 114   B      0.0847     1
8 115   B     -1.05       2

One of the suggested solutions for similar questions is

library(dplyr)
df %>% group_by(s1) %>% mutate(id = row_number())

but that resets each time s1 changes. Similarly, these did not work either:

df %>% group_by(s1, s2) %>% mutate(id = row_number())
df %>% group_by(s2) %>% mutate(id = row_number())
df %>% group_by(s1) %>% mutate(id = row_number(s2))
df %>% group_by(s1) %>% mutate(id = cur_group_id())
df %>% group_by(s1, s2) %>% mutate(id = cur_group_id())