[bug report] firmware: arm_scmi: Call Raw mode hooks from the core stack

[Dan Carpenter]

On Wed, Jan 18, 2023 at 04:28:38PM +0000, Cristian Marussi wrote:
> On Wed, Jan 18, 2023 at 06:41:19PM +0300, Dan Carpenter wrote:
> > On Tue, Jan 17, 2023 at 02:14:30PM +0000, Cristian Marussi wrote:
> > > 
> > > I'm still puzzled why I regularly do not catch these errors on my smatch
> > > setup running kchecker which does only reports me a bunch of errors on
> > > core bitops code. (I have recently pulled updated smatch branch...)
> > 
> > This check requires the cross function database to work.  Building the
> > cross function database is easy enough, but takes overnight.
> > 
> > ~/smatch/smatch_scripts/build_kernel_data.sh
> > 
> 
> Thanks I'll try.
> 
> BTW I'm using v1.73 and I also see since a while a lot of errors like:
> 
> internal error: file too large:  db/kernel.return_fixes (limit 4096 bytes)
> 

That's an interesting thing.  I had to make db/kernel.return_fixes
larger a few months back, but why is the db/kernel.return_fixes out of
sync with the code to parse it in v1.73?

> internal error: dev/src/pdsw/linux/arch/arm64/include/asm/jump_label.h:38 SQL error #2: UNIQUE constraint failed: mtag_data.tag, mtag_data.offset, mtag_data.type, mtag_data.value
> 
> internal error: dev/src/pdsw/linux/arch/arm64/include/asm/jump_label.h:38 SQL: 'insert into mtag_data values (7179031484371304448, 0, 8041, '1:      b               %l[l_yes]               \n.pushsection     __jump_table, "aw"      \n              .align          3                       \n              .long           1b - ., %l[l_yes] - .   \n              .quad           %c0 - .                 \n.popsection                              \n      ');'
> 
> Trying to revert back a few months of commits and those errors disappear.
> Is it related to have stale DB ?

Yep.  Probably.

Can you retry on the latest git?

regards,
dan carpenter