Re: [PATCH 3/3] powerpc: icswx: Simple ACOP fault handler for both book3e and book3s parts.

2011-08-09 Thread Jimi Xenidis
On Aug 9, 2011, at 12:26 AM, Kumar Gala wrote: On Aug 8, 2011, at 5:26 PM, Jimi Xenidis wrote: This patch adds a fault handler that responds to illegal Coprocessor types. Currently all CTs are treated and illegal. There are two ways to report the fault back to the application. If the

Re: [PATCH 3/3] powerpc: icswx: Simple ACOP fault handler for both book3e and book3s parts.

2011-08-09 Thread Benjamin Herrenschmidt
On Tue, 2011-08-09 at 00:26 -0500, Kumar Gala wrote: + /* Some implementations leave us a hint for the CT */ + ct = ICSWX_GET_CT_HINT(error_code); + if (ct 0) { + /* we have to peek at the instruction work to figure out CT */ + union cop_ccw ccw; don't use

Re: [PATCH 3/3] powerpc: icswx: Simple ACOP fault handler for both book3e and book3s parts.

2011-08-09 Thread Jimi Xenidis
On Aug 9, 2011, at 10:15 AM, Benjamin Herrenschmidt wrote: On Tue, 2011-08-09 at 00:26 -0500, Kumar Gala wrote: + /* Some implementations leave us a hint for the CT */ + ct = ICSWX_GET_CT_HINT(error_code); + if (ct 0) { + /* we have to peek at the instruction work to

[PATCH 3/3] powerpc: icswx: Simple ACOP fault handler for both book3e and book3s parts.

2011-08-08 Thread Jimi Xenidis
This patch adds a fault handler that responds to illegal Coprocessor types. Currently all CTs are treated and illegal. There are two ways to report the fault back to the application. If the application used the record form (icswx.) then the architected reject is emulated. If the application did

Re: [PATCH 3/3] powerpc: icswx: Simple ACOP fault handler for both book3e and book3s parts.

2011-08-08 Thread Kumar Gala
On Aug 8, 2011, at 5:26 PM, Jimi Xenidis wrote: This patch adds a fault handler that responds to illegal Coprocessor types. Currently all CTs are treated and illegal. There are two ways to report the fault back to the application. If the application used the record form (icswx.) then the