Re: [PATCH] bcachefs: don't call sleeping funcs when handling inconsistency errors

[Kent Overstreet]

On Fri, Apr 04, 2025 at 12:22:45PM +0530, Bharadwaj Raju wrote:
> On Wed, Apr 2, 2025 at 10:08 PM Kent Overstreet
> <kent.overstreet@xxxxxxxxx> wrote:
> >
> > On Wed, Apr 02, 2025 at 10:03:10PM +0530, Bharadwaj Raju wrote:
> > > On Wed, Apr 2, 2025 at 9:47 PM Kent Overstreet
> > > <kent.overstreet@xxxxxxxxx> wrote:
> >
> > If you're feeling particularly adventurous - print_string_as_lines() is
> > a hack, I think we should be able to do something more robust by
> > skipping printk (that's where the 1k limit comes from) and calling
> > something lower level - that will require digging into the printk
> > codepath and finding lower level we can call.
> 
> I tried looking into the printk codepath, namely vprintk_emit -> vprintk_store.
> It doesn't seem like there's a convenient single lower-level
> entrypoint we could
> call which just avoids the 1k limit, rather there's a lot of internal
> code mixed with
> the truncation that we'd have to just copy if we want printk behavior.
> I don't think that's a reasonable option for us.

Yeah that does look a bit messy - and it doesn't exactly explain where
the 1k limit comes in.


> > I also just noticed that print_string_as_lines() needs to check for
> > being passed a NULL pointer - in case the printbuf memory allocation
> > fails. Want to get that one too?
> 
> From what I'm seeing __bch2_print_string_as_lines already checks
> for the lines pointer being NULL. The only unchecked pointer is prefix,
> which I don't think needs to be checked since it will be something constant
> from kern_levels.h, not something which could be NULL.

Yep, I should've checked the code before I said that :)