Re: [PATCH 09/11] include/exec: added functions to the stubs in exec-all.h

From: David Gibson <david@gibson.dropbear.id.au>
To: Bruno Piazera Larsen <bruno.larsen@eldorado.org.br>
Cc: farosas@linux.ibm.com,
	Richard Henderson <richard.henderson@linaro.org>,
	qemu-devel@nongnu.org,
	Lucas Mateus Martins Araujo e Castro
	<lucas.araujo@eldorado.org.br>,
	fernando.valle@eldorado.org.br, qemu-ppc@nongnu.org,
	matheus.ferst@eldorado.org.br, luis.pires@eldorado.org.br
Subject: Re: [PATCH 09/11] include/exec: added functions to the stubs in exec-all.h
Date: Mon, 24 May 2021 13:15:34 +1000	[thread overview]
Message-ID: <YKsaVk/3Ixd64g+y@yekko> (raw)
In-Reply-To: <92660b00-e2a4-056b-0a66-e729f502cc19@eldorado.org.br>

[-- Attachment #1: Type: text/plain, Size: 2598 bytes --]

On Mon, May 17, 2021 at 08:07:24AM -0300, Bruno Piazera Larsen wrote:
> 
> On 17/05/2021 00:55, David Gibson wrote:
> > On Thu, May 13, 2021 at 11:03:24AM -0300, Lucas Mateus Martins Araujo e Castro wrote:
> > > On 12/05/2021 15:34, Richard Henderson wrote:
> > > > On 5/12/21 9:08 AM, Bruno Larsen (billionai) wrote:
> > > > > From: "Lucas Mateus Castro (alqotel)"<lucas.araujo@eldorado.org.br>
> > > > > 
> > > > > Added tlb_set_page and tlb_set_page_with_attrs to the
> > > > > stubbed functions in exec-all.h  as it is needed
> > > > > in some functions when compiling without TCG
> > > > > 
> > > > > Signed-off-by: Lucas Mateus Castro
> > > > > (alqotel)<lucas.araujo@eldorado.org.br>
> > > > > ---
> > > > >    include/exec/exec-all.h | 10 ++++++++++
> > > > >    1 file changed, 10 insertions(+)
> > > > No, the caller is tcg-specific already.
> > > > 
> > > > 
> > > > r~
> > > tlb_set_page is called by many ppc_hash64_handle_mmu_fault,
> > > ppc_radix64_handle_mmu_fault and ppc_hash32_handle_mmu_fault, all of which
> > > from what I've seen are only used inside #if defined(CONFIG_SOFTMMU). So
> > > what is the best way to deal with these tlb_set_page calls? Should these
> > > part of the _handle_mmu_fault functions never be reached or should
> > > these
> > The handle_mmu_fault() functions per se shouldn't be included in a
> > !SOFTMMU build.  We might have to extract some of their internal logic
> > for the gdb path, though.
> > 
> > > functions never be called?
> > > 
> > > If it's the latter then should we change the #if defined to #if
> > > defined(CONFIG_SOFTMMU) && (CONFIG_TCG)?
> > That definitely doesn't make sense.  In practice CONFIG_SOFTMMU == CONFIG_TCG.

Ugh.. sorry.. I was confused again.  In practice CONFIG_SOFTMMU ==
!CONFIG_USER_ONLY.

> We figured it was the case, but from what I can tell, CONFIG_SOFTMMU is set
> when parsing the target list (in the configure script) and CONFIG_TCG is set
> later, when parsing which accelerators were requested. So even though
> SOFTMMU should imply TCG, the way it is coded right now doesn't. We could
> also try and change the configure script, but neither of us is really good
> with bash scripts, so this was the next best solution we came up with.
> > 
> > > 
> > > P.S: There was a miscommunication between me and Bruno, this should've been
> > > a RFC.
> > > 

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]