Current guest kernels try allocating as many vectors as the quota is.
For example, in the case of virtio-net (which has just 3 vectors)
the guest requests 4 vectors (that is the quota in the test) and
the existing ibm,change-msi handler returns 4. But before it returns,
it calls msix_set_message() in a loop and corrupts memory behind
the end of msix_table.
This limits the number of vectors returned by ibm,change-msi to
the maximum supported by the actual device.
Signed-off-by: Alexey Kardashevskiy <[email protected]>
Cc: [email protected]
[agraf: squash in bugfix from aik]
Signed-off-by: Alexander Graf <[email protected]>
/* There is no cached config, allocate MSIs */
if (!phb->msi_table[ndev].nvec) {
+ int max_irqs = 0;
+ if (ret_intr_type == RTAS_TYPE_MSI) {
+ max_irqs = msi_nr_vectors_allocated(pdev);
+ } else if (ret_intr_type == RTAS_TYPE_MSIX) {
+ max_irqs = pdev->msix_entries_nr;
+ }
+ if (!max_irqs) {
+ error_report("Requested interrupt type %d is not enabled for device#%d",
+ ret_intr_type, ndev);
+ rtas_st(rets, 0, -1); /* Hardware error */
+ return;
+ }
+ if (req_num > max_irqs) {
+ req_num = max_irqs;
+ }
irq = spapr_allocate_irq_block(req_num, false,
ret_intr_type == RTAS_TYPE_MSI);
if (irq < 0) {