Re: [PATCH v5 06/21] KVM: selftests: Add helper to get host IRQ from device MSIX for IRQ bypass test

[Sean Christopherson <seanjc@google.com>]

On Thu, Jun 04, 2026, sashiko-bot@kernel.org wrote:
> > +unsigned int vfio_msix_to_host_irq(const char *device_bdf, int msix)
> > +{
> > +	char search_string[64];
> > +	char line[4096];
> > +	int irq = -1;
> > +	FILE *fp;
> > +
> > +	fp = open_proc_interrupts();
> > +
> > +	snprintf(search_string, sizeof(search_string), "vfio-msix[%d]", msix);
> > +
> > +	while (fgets(line, sizeof(line), fp)) {
> > +		if (strstr(line, device_bdf) && strstr(line, search_string)) {
> > +			TEST_ASSERT_EQ(1, sscanf(line, "%d:", &irq));
> 
> [Severity: Medium]
> Is the 4096-byte buffer large enough for systems with high core counts?
> 
> In /proc/interrupts, each online CPU gets its own column for interrupt counts.
> On systems with many CPUs, a single line might exceed 4096 bytes. When
> this happens, fgets() reads the line in multiple chunks.
> 
> If the line is chunked, the first chunk contains the IRQ number, and a
> subsequent chunk contains the target device names. The strstr() condition
> would match on the later chunk, but sscanf() would look for the IRQ number
> format at the beginning of that same chunk.
> 
> Since the later chunk lacks the leading IRQ number, sscanf() would return 0,
> which would then cause TEST_ASSERT_EQ to fail and crash the test.

I'm fine punting on this for now.  Crashing is unfortunate, but easy enough to
workaround if someone with a massive system wants to run the test, and more
importantly shouldn't mislead the user.