The new legacy_mem file in sysfs is causing problems with X on machines
that don't support legacy memory access. The way I initially implemented
it, we would fail with -ENXIO when trying to mmap it, thus exposing to
X that we do support the API but there is no legacy memory.
Unfortunately, X poor error handling is causing it to fail to start when
it gets this error.
This implements a workaround hack that instead maps anonymous memory
instead (using shmem if VM_SHARED is set, just like /dev/zero does).
Signed-off-by: Benjamin Herrenschmidt <b...@kernel.crashing.org>
---
This fixes a regression in .29 and is the less ugly way we found to
sort this problem out. I'll merge it ASAP so holler quick if you
have an objection
arch/powerpc/kernel/pci-common.c | 17 +++++++++++++++--
1 file changed, 15 insertions(+), 2 deletions(-)
--- linux-work.orig/arch/powerpc/kernel/pci-common.c 2009-02-06
16:23:36.000000000 +1100
+++ linux-work/arch/powerpc/kernel/pci-common.c 2009-02-06 16:30:32.000000000
+1100
@@ -561,8 +561,21 @@ int pci_mmap_legacy_page_range(struct pc
(unsigned long long)(offset + size - 1));
if (mmap_state == pci_mmap_mem) {
- if ((offset + size) > hose->isa_mem_size)
- return -ENXIO;
+ /* Hack alert !
+ *
+ * Because X is lame and can fail starting if it gets an error
trying
+ * to mmap legacy_mem (instead of just moving on without legacy
memory
+ * access) we fake it here by giving it anonymous memory,
effectively
+ * behaving just like /dev/zero
+ */
+ if ((offset + size) > hose->isa_mem_size) {
+ printk(KERN_DEBUG
+ "Process %s (pid:%d) mapped non-existing PCI
legacy memory for 0%04x:%02x\n",
+ current->comm, current->pid, pci_domain_nr(bus),
bus->number);
+ if (vma->vm_flags & VM_SHARED)
+ return shmem_zero_setup(vma);
+ return 0;
+ }
offset += hose->isa_mem_phys;
} else {
unsigned long io_offset = (unsigned long)hose->io_base_virt -
_IO_BASE;
_______________________________________________
Linuxppc-dev mailing list
Linuxppc-dev@ozlabs.org
https://ozlabs.org/mailman/listinfo/linuxppc-dev
