Re: [edk2-devel] [PATCH] RedfishPkg/RedfishRestExDxe: return HTTP status code to caller.

["Nickle Wang via groups.io" <nicklew=nvidia.com@groups.io>]

Hi Make,

> I would suggest to remove this `if (EFI_ERROR (Status)) {` block at all.

Are you suggesting to not release "ResponseMessage->Body"[1] while error happens because caller will handle this too?

[1]: Line 452 - 454 at: https://github.com/tianocore/edk2/pull/4672/files#diff-18a31a76c495b9a145b0ffa6a68fe5de887cbdddc925c5eb3a759e81694f59c9L452

Thanks,
Nickle

> -----Original Message-----
> From: Mike Maslenkin <mike.maslenkin@gmail.com>
> Sent: Saturday, July 22, 2023 8:09 PM
> To: devel@edk2.groups.io; Nickle Wang <nicklew@nvidia.com>
> Cc: Abner Chang <abner.chang@amd.com>; Igor Kulchytskyy <igork@ami.com>;
> Nick Ramirez <nramirez@nvidia.com>
> Subject: Re: [edk2-devel] [PATCH] RedfishPkg/RedfishRestExDxe: return HTTP
> status code to caller.
> 
> External email: Use caution opening links or attachments
> 
> 
> On Sat, Jul 22, 2023 at 11:18 AM Nickle Wang via groups.io
> <nicklew=nvidia.com@groups.io> wrote:
> >
> > Return unsupported HTTP status code to caller so caller can handle
> > HTTP error status code. Current implementation only return EFI error
> > to caller. Without knowing the HTTP status code, caller has trouble to
> > handle HTTP request failure.
> >
> > Signed-off-by: Nickle Wang <nicklew@nvidia.com>
> > Cc: Abner Chang <abner.chang@amd.com>
> > Cc: Igor Kulchytskyy <igork@ami.com>
> > Cc: Nick Ramirez <nramirez@nvidia.com>
> > ---
> >  .../RedfishRestExDxe/RedfishRestExProtocol.c    | 17 ++++++++++++-----
> >  1 file changed, 12 insertions(+), 5 deletions(-)
> >
> > diff --git a/RedfishPkg/RedfishRestExDxe/RedfishRestExProtocol.c
> > b/RedfishPkg/RedfishRestExDxe/RedfishRestExProtocol.c
> > index 90973619f2bc..f11cee8542fb 100644
> > --- a/RedfishPkg/RedfishRestExDxe/RedfishRestExProtocol.c
> > +++ b/RedfishPkg/RedfishRestExDxe/RedfishRestExProtocol.c
> > @@ -320,6 +320,18 @@ ReSendRequest:;
> >      DEBUG ((DEBUG_ERROR, "This HTTP Status is not handled!\n"));
> >      DumpHttpStatusCode (DEBUG_REDFISH_NETWORK, ResponseData-
> >Response.StatusCode);
> >      Status = EFI_UNSUPPORTED;
> > +
> > +    //
> > +    // Deliver status code back to caller so caller can handle it.
> > +    //
> > +    ResponseMessage->Data.Response = AllocateZeroPool (sizeof
> (EFI_HTTP_RESPONSE_DATA));
> > +    if (ResponseMessage->Data.Response == NULL) {
> > +      Status = EFI_OUT_OF_RESOURCES;
> > +      goto ON_EXIT;
> > +    }
> > +
> > +    ResponseMessage->Data.Response->StatusCode =
> > + ResponseData->Response.StatusCode;
> > +
> >      goto ON_EXIT;
> >    }
> >
> > @@ -444,11 +456,6 @@ ON_EXIT:
> >    }
> >
> >    if (EFI_ERROR (Status)) {
> > -    if (ResponseMessage->Data.Response != NULL) {
> > -      FreePool (ResponseMessage->Data.Response);
> > -      ResponseMessage->Data.Response = NULL;
> > -    }
> > -
> This doesn't make code cleaner.
> What about resource deallocation after  AllocateZeroPool (sizeof
> (EFI_HTTP_RESPONSE_DATA)) for the common case?
> I guess this resource deallocation was added for this and there are a lot of 'goto
> ON_EXIT;' after original AllocateZeroPool call.
> I would suggest to remove this `if (EFI_ERROR (Status)) {` block at all.
> As I can see, the only callers of this function are
> RedfishPkg//PrivateLibrary/RedfishLib/edk2libredfish/src/service.c and edk2-
> redfish-
> client//RedfishClientPkg/PrivateLibrary/RedfishLib/edk2libredfish/src/service.c.
> And both use the same pattern of an explicit response data deallocation
> (RestConfigFreeHttpMessage() call)  in case of error status returned.
> I.e. caller is responsible for the response data deallocation as it owns it.


-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#107517): https://edk2.groups.io/g/devel/message/107517
Mute This Topic: https://groups.io/mt/100292372/7686176
Group Owner: devel+owner@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [rebecca@openfw.io]
-=-=-=-=-=-=-=-=-=-=-=-